TILL DEATH DO US PART: PREPUBLICATION REVIEW IN
THE INTELLIGENCE COMMUNITY
Kevin Casey*
As a condition of access to classiﬁed information, most employees of
the U.S. intelligence community are required to sign nondisclosure
agreements that mandate lifetime prepublication review. In essence,
these agreements require employees to submit any works that discuss
their experiences working in the intelligence community---whether written or oral, ﬁction or nonﬁction---to their respective agencies and receive
approval before seeking publication. Though these agreements constitute
an exercise of prior restraint, the Supreme Court has held them constitutional. This Note does not argue for or against the constitutionality of
prepublication review; instead, it explores how prepublication review is
actually practiced by agencies and concludes that the current system,
which lacks executive-branch-wide guidance, grants too much discretion
to individual agencies. It compares the policies of individual agencies
with the experiences of actual authors who have clashed with
prepublication-review boards to argue that agencies conduct review in a
manner that is inconsistent at best, and downright biased and
discriminatory at worst.
The level of secrecy shrouding intelligence agencies and the
concomitant dearth of publicly available information about their activities make it difficult to evaluate their performance and, by extension,
the performance of our elected officials in overseeing such activities. In
such circumstances, memoirs and other forms of expression by former
agency employees become extremely valuable. The potential for
discriminatory review---the approval of works that portray agencies in a
positive light and the suppression of works more critical in tone--illuminates the need for an improved system of prepublication review:
one that respects the intelligence community’s need to protect legitimate
national-security information but demands more robust protections for
the First Amendment rights of potential authors and the public’s need
for information with which to evaluate the highly secretive activities of
their government. This Note concludes by arguing that action is required from all three branches of government to improve the system of
prepublication review.
INTRODUCTION
In 2009, Anthony Shaffer, a retired Lieutenant Colonel and career
intelligence officer, sought to publish a memoir depicting his experi*. J.D. Candidate 2015, Columbia Law School.
417
418
COLUMBIA LAW REVIEW
[Vol. 115:417
ences working with the Defense Intelligence Agency (DIA) as an Army
Reserve officer after 9/11. As required by several nondisclosure agreements he had signed to gain access to classiﬁed information, Shaffer
submitted his manuscript to his superiors in the Army for prepublication
review. After review by “two highly qualiﬁed Army Reserve officers,”1 the
Army determined the manuscript contained no classiﬁed information
and approved it for publication. Shaffer forwarded the manuscript to his
publishers. Shortly thereafter, the DIA intervened, demanding a copy of
the manuscript so it could conduct its own review. The Army complied;
meanwhile, the publisher continued working toward publication. Three
weeks before the scheduled shipment of the ﬁrst edition, the DIA notiﬁed Shaffer that the Army’s review process was insufficient and that he
would have to submit his manuscript officially to the DIA for proper
review. Shaffer agreed to comply, but his publisher notiﬁed the government that it had already sent out several dozen copies of the book to
reviewers. Undeterred, the government pressed forward with its demands
for redactions and paid the publisher nearly $50,000 to destroy all
remaining copies of the ﬁrst edition.2 The realities of the Internet Age,
however, brought about a farcical result—various news organizations and
private citizens purchased the unredacted ﬁrst edition and published
side-by-side comparisons of the two versions,3 revealing not just the allegedly classiﬁed information, but what types of information the government considered sensitive enough to require redaction—itself a harm
that the government seeks to avoid.4 As discussed throughout this Note,
Shaffer’s experience is not unique.
Employees of intelligence community (IC) agencies5 are required to
sign a nondisclosure agreement that incurs certain lifetime obligations.6
Two of the obligations are unsurprising: The employee agrees not to disclose classiﬁed information—information that is properly classiﬁed
1. First Amended Complaint at 5, Shaffer v. Def. Intelligence Agency, 901 F. Supp.
2d 113 (D.D.C. 2013) (No. 10-2119), available at http://www.fas.org/sgp/jud/shaffer/021
312-complaint.pdf (on ﬁle with the Columbia Law Review).
2. Id. at 11.
3. Id. at 12–14.
4. See Def. Intelligence Agency, DIA Form No. 271, Conditions of Employment 5
(Nov. 20, 2008) [hereinafter DIA Employment Conditions], available at http://www.dia.
mil/Portals/27/Documents/Careers/Pre-employment%20Forms/ConditionsofEmployme
nt.doc (on ﬁle with the Columbia Law Review) (“An agreement is also required to authorize
Agency pre-publication review of certain material prior to disclosure during and after
employment with the DIA.”).
5. See infra notes 13–14 and accompanying text (describing background and
activities of IC).
6. See Exec. Order No. 12,968, 3 C.F.R. 391, 392 (1996) (requiring signing of
nondisclosure agreement as precondition to accessing classiﬁed information); 32 C.F.R.
§ 2001.80 (2014) (requiring agencies to either use Standard Form (SF) 312 as their
nondisclosure agreement or apply for waiver).
2015]
PREPUBLICATION REVIEW
419
under Executive Order 13,5267—and to report to the Federal Bureau of
Investigation (FBI) any attempt made by unauthorized persons to solicit
classiﬁed information.8 But in most cases, a third obligation is incurred:
The employee must submit any works based on her experiences in the
IC, whether ﬁctional or nonﬁctional, for prepublication review. 9
Importantly, this is a lifelong commitment that remains valid even after
separation from the agency.
The Supreme Court found such a lifetime obligation constitutional
in Snepp v. United States.10 Much of the legal literature on prepublication
review has focused on its constitutionality, scrutinizing the Court’s decision in that case. 11 This Note does not argue for or against the
constitutionality of prepublication review. Instead, it attempts to shed
light on the disarray of the current system through analysis of disparate
agency policies and the experiences of authors pursuing publication. It
concludes that the current system—conducted by individual agencies in
a decentralized fashion, with no executive-branch-wide guidance—leads
to arbitrary, inconsistent, and sometimes absurd results and does not conform with controlling case law, particularly with respect to former employees. It also deprives the public of important insight into a highly secretive
area of government. The system is in need of an overhaul, and this Note
proposes several avenues for reform.
Part I discusses the history of prepublication review and its
constitutionality, decided under controversial circumstances in Snepp. It
also analyzes United States v. Marchetti,12 a Fourth Circuit decision upholding prepublication review that predates Snepp but has been relied upon
as authoritative guidance both for agencies conducting prepublication
review and for potential authors challenging it. Part II illuminates the
prepublication-review policies of the IC agencies based on publicized and
declassiﬁed guidelines and compares them with the actual experiences of
current and former employees who have sought to have their memoirs
published. It concludes that the unusually broad discretion granted to
agencies in determining their prepublication-review policies creates a
risk of abuse that is particularly problematic in light of the authors’ First
7. Information cannot be classiﬁed unless its unauthorized disclosure “could
reasonably be expected to cause identiﬁable or describable damage to national security”
and it pertains to one of eight enumerated categories. See Exec. Order No. 13,526, 3
C.F.R. 298, 300 (2010) (listing categories).
8. U.S. Gen. Servs. Admin., Standard Form No. 312, Classiﬁed Information Nondisclosure Agreement (July 2013) [hereinafter Standard Form 312], available at http://
www.archives.gov/isoo/security-forms/sf312.pdf (on ﬁle with the Columbia Law Review).
9. See infra notes 100–102 and accompanying text (describing agency prepublication-review requirements).
10. 444 U.S. 507 (1980) (per curiam). For a discussion of the case, see infra Part
I.B.3–4.
11. See infra notes 80–81 and accompanying text (listing articles considering constitutionality of review).
12. 466 F.2d 1309 (4th Cir.), cert. denied, 409 U.S. 1063 (1972).
420
COLUMBIA LAW REVIEW
[Vol. 115:417
Amendment rights. Finally, Part III proposes several solutions, arguing
action is required from all three branches of government.
I. THE HISTORY AND CONSTITUTIONALITY OF PREPUBLICATION REVIEW
The history and constitutional rationale of prepublication review is
important in understanding the role it plays in an environment of
unprecedented secrecy. Part I discusses the history and constitutionality
of prepublication-review policies. Part I.A seeks to demonstrate the value
that publications by IC employees add to the dialogue on nationalsecurity issues. Part I.A.1 discusses the inherent advantages that the
executive branch has in controlling that discourse, while Part I.A.2 and
I.A.3 explore how such publications can counter this advantage. Part I.B
focuses on the constitutionality of prepublication review, with Part I.B.1
ﬁrst contextualizing review policies by discussing the First Amendment’s
strong presumption against prior restraint. Part I.B.2 analyzes the Fourth
Circuit’s decision in Marchetti, while Part I.B.3 discusses the Supreme
Court’s decision in Snepp. Part I.B.4 explores the aftermath of Snepp and
the lasting impact the decision has had on prepublication-review policies.
A.
The Value of Publications by Current and Former Employees
1. The Executive’s Advantage. — The IC is a “federation of executive
branch agencies and organizations that work separately and together to
conduct intelligence activities necessary for the conduct of foreign relations and the protection of the national security of the United States.”13
Its activities include the “[c]ollection of information needed by the
President, the National Security Council, the Secretaries of State and
Defense, and other executive-branch officials for the performance of
their duties and responsibilities.” 14 The breadth of its activities has
increased dramatically in recent years,15 as reﬂected in the rate at which
its budget has grown.16 The unauthorized disclosure of highly classiﬁed
13. Office of the Dir. of Nat’l Intelligence, ODNI FAQ 7, http://www.dni.gov/index.
php/about/faq?tmpl=component&format=pdf (on ﬁle with the Columbia Law Review) (last
visited Nov. 6, 2014).
14. Id.
15. See, e.g., Dana Priest & William M. Arkin, A Hidden World, Growing Beyond
Control, Wash. Post (July 19, 2010, 4:50 PM), http://projects.washingtonpost.com/top-secr
et-america/articles/a-hidden-world-growing-beyond-control/print/ (on ﬁle with the
Columbia Law Review) (describing rapid growth in number of personnel employed, and
volume of data analyzed, by various agencies); Scott Shane, No Morsel Too Minuscule for
All-Consuming N.S.A., N.Y. Times (Nov. 2, 2013), http://www.nytimes.com/2013/11/03/
world/no-morsel-too-minuscule-for-all-consuming-nsa.html?pagewanted=all (on ﬁle with
the Columbia Law Review) (describing National Security Agency’s expansive collection of
communications data).
16. The National Intelligence Program budget, which funds all intelligence activities
except those under the purview of the military, for ﬁscal year 2014 was $50.5 billion. Press
Release, Office of the Dir. of Nat’l Intelligence, DNI Releases Budget Figure for FY 2014
National Intelligence Program (Oct. 30, 2014), http://www.dni.gov/index.php/newsroo
2015]
PREPUBLICATION REVIEW
421
information by former National Security Agency (NSA) contractor
Edward Snowden has shed some light on the activities of the IC,17 leading
to calls for more official disclosure.18
Yet despite its increasingly inﬂuential role, the IC remains highly
secretive,19 and the courts have adopted a strong presumption in support
of secrecy. The Supreme Court has declared that “no governmental interest is more compelling than the security of the Nation” 20 and has
explicitly affirmed the President’s constitutional power to classify
information for its protection. 21 The Court has also recognized the
government’s “compelling interest” in withholding national-security
information from unauthorized persons in the course of executive business.22 The doctrine of executive privilege shields some executive-branch
secrets from congressional scrutiny,23 and courts have been hesitant to
m/press-releases/198-press-releases-2014/1134-dni-releases-budget-ﬁgure-for-fy-2014-natio
nal-intelligence-program (on ﬁle with the Columbia Law Review). The Military Intelligence
Program budget for the same period was $17.4 billion. Press Release, Dep’t of Def., DOD
Releases Military Intelligence Program (MIP) Appropriated Top Line Budget for Fiscal
Year (FY) 2014 (Oct. 30, 2014), http://www.defense.gov/releases/release.aspx?releaseid=
17010 (on ﬁle with the Columbia Law Review). The total comes to $67.9 billion. This
reﬂects a marked increase since 9/11. See Barton Gellman & Greg Miller, ‘Black Budget’
Summary Details U.S. Spy Network’s Successes, Failures and Objectives, Wash. Post (Aug.
29, 2013), http://www.washingtonpost.com/world/national-security/black-budget-summa
ry-details-us-spy-networks-successes-failures-and-objectives/2013/08/29/7e57bb78-10ab-1
1e3-8cdd-bcdc09410972_story.html (on ﬁle with the Columbia Law Review) (reporting
massive increases in budgets, including one-hundred percent increase in IC budget, since
2001).
17. See, e.g., Glenn Greenwald, XKeyscore: NSA Tool Collects ‘Nearly Everything a
User Does on the Internet,’ Guardian (July 31, 2013, 8:56 AM), http://www.theguardian.
com/world/2013/jul/31/nsa-top-secret-program-online-data (on ﬁle with the Columbia
Law Review) (describing NSA collection of online data).
18. The Foreign Intelligence Surveillance Court recently ordered the Department of
Justice to release secret opinions about section 215 of the Patriot Act. In re Orders of This
Court Interpreting Section 215 of the Patriot Act, No. Misc. 13-02, slip op. at 16 (FISA Ct.
Sept. 13, 2013), available at http://www.uscourts.gov/uscourts/courts/ﬁsc/misc-13-02order-130813.pdf (on ﬁle with the Columbia Law Review) (“[U]nauthorized disclosure in
June 2013 of a Section 215 order, and government statements in response to that
disclosure, have engendered considerable public interest and debate about Section 215.
Publication of FISC opinions relating to this provision would contribute to an informed
debate.”).
19. See Info. Sec. Oversight Office, Nat’l Archives & Records Admin., 2013 Report to
the President 1 (2013) [hereinafter 2013 ISOO Report], available at http://www.archives.
gov/isoo/reports/2013-annual-report.pdf (on ﬁle with the Columbia Law Review) (reporting over eighty million classiﬁcation decisions in 2013).
20. Haig v. Agee, 453 U.S. 280, 307 (1981).
21. See Dep’t of the Navy v. Egan, 484 U.S. 518, 527 (1988) (“[The President’s]
authority to classify and control access to information bearing on national security . . .
ﬂows primarily from this constitutional investment of power in the President and exists
quite apart from any explicit congressional grant.”).
22. Id. (quoting Snepp v. United States, 444 U.S. 507, 509 n.3 (1980) (per curiam)).
23. See Todd Garvey & Alissa M. Dolan, Cong. Research Serv., R42670, Presidential
Claims of Executive Privilege: History, Law, Practice, and Recent Developments 8 (2012),
422
COLUMBIA LAW REVIEW
[Vol. 115:417
scrutinize national-security matters themselves.24 The absence of checks
and balances from the legislative and judicial branches of government
has led to an accountability and oversight issue unlike any other. In other
areas where these two branches exercise deference, voters act as the primary check on overzealous executive action;25 but because the activities
of IC agencies are so secretive, the public is unable to make informed
evaluations about their conduct and efficacy.26
In contrast, there are many opportunities for the executive branch
to portray itself in a favorable light. One prevalent method is the use of
strategic leaks of information to the media.27 This practice is common in
Washington 28 and has been so for at least half a century. 29 Another
method is direct inﬂuence, criticized most recently during the controversy surrounding the release of the motion picture Zero Dark Thirty. A
declassiﬁed memo indicates the Central Intelligence Agency (CIA) was
involved in the movie’s production in order to “help promote an
appropriate portrayal of the Agency and the Bin Laden operation.”30 A
available at http://fas.org/sgp/crs/secrecy/R42670.pdf (on ﬁle with the Columbia Law
Review) (reporting executive-branch statements have identiﬁed foreign relations and
military affairs as presumptively covered by executive privilege).
24. See, e.g., Stephen Dycus et al., National Security Law 124 (5th ed. 2011) (“Courts
have . . . tended to avoid the decision of national security disputes and thus to defer to the
political branches in what may be an increasing number of cases.”).
25. See Morrison v. Olson, 487 U.S. 654, 711 (1988) (Scalia, J., dissenting) (arguing
“checks against any branch’s abuse of its exclusive power” include “political check that the
people will replace those in the political branches . . . who are guilty of abuse”).
26. See 1 Arvin S. Quist, Security Classiﬁcation of Information 138 (rev. Sept. 20,
2002), http://fas.org/sgp/library/quist/chap_6.pdf (on ﬁle with the Columbia Law Review)
(“When information is classiﬁed, the public’s knowledge of the government’s activities is
reduced, thereby impeding an informed public evaluation of governmental policies and
government officials.”).
27. See generally David E. Pozen, The Leaky Leviathan: Why the Government
Condemns and Condones Unlawful Disclosures of Information, 127 Harv. L. Rev. 512
(2013) [hereinafter Pozen, Leaky Leviathan] (discussing beneﬁt executive branch gains
through current system of plants, leaks, and “pleaks”); Robert A. Sedler, The Media and
National Security, 53 Wayne L. Rev. 1025, 1034 (2007) (“Sometimes government officials,
in order to advance the government’s purpose, voluntarily disclose information to the
media, so that the media will assist them in conveying the government’s message to the
public.”).
28. See Pozen, Leaky Leviathan, supra note 27, at 528 (“It is a commonplace that
leaks course through the nation’s capital. Classiﬁed information disclosures to the media
are thought to occur so regularly in Washington as to constitute a routine method of
communication about government.” (quoting William E. Lee, Deep Background:
Journalists, Sources, and the Perils of Leaking, 57 Am. U. L. Rev. 1453, 1467 (2008))
(internal quotation marks omitted)).
29. See Affidavit of Max Frankel para. 17, United States v. N.Y. Times Co., 328 F.
Supp. 324 (S.D.N.Y. 1971) (No. 71 Civ. 2662), 1971 WL 224067 (describing “informal but
customary traffic in secret information” among reporters and officials in Washington).
30. Redacted Internal CIA Memorandum 1 (approved for release Apr. 22, 2013),
available at https://s3.amazonaws.com/s3.documentcloud.org/documents/696468/boalcia-memo.pdf (on ﬁle with the Columbia Law Review).
2015]
PREPUBLICATION REVIEW
423
subsequent inquiry requested by Congress focused on whether classiﬁed
information was provided; the fact that the Administration was involved
in the ﬁlmmaking process was treated as almost de rigueur.31
2. Publications by Insiders Offer Alternatives to Official Narrative. — The
abundance of authorized—or at least condoned—leaks combined with
the paucity of information available from other sources makes it difficult
for the public to evaluate the work of the IC effectively. In this context,
memoirs and speeches by current and former employees provide the
public with unique insight: The authors are insiders who have ﬁrsthand
experience of the inner workings of the IC. Undoubtedly, many will be
agency loyalists who provide nothing unavailable through official leaks
and releases,32 but the works of critical or neutral authors can provide the
public with an otherwise unavailable perspective into the inner workings
of secretive organizations.33
For example, the views of insiders have provided critical insight into
the debate over the use of so-called “enhanced interrogation techniques.”34 Former senior CIA officials have defended the program, insisting the techniques led to actionable intelligence. 35 Former Vice
President Dick Cheney has also spoken approvingly of such techniques,
31. There were allegations that CIA Director Leon Panetta disclosed classiﬁed
information to the directors, but the Department of Defense (DoD) Inspector General’s
(IG) ﬁnal report states that both the DoD and CIA took all appropriate measures to
protect classiﬁed information during their interactions with the movie producers. Office
of the Deputy Inspector Gen. for Intelligence & Special Program Assessments, Dep’t of
Def., Report No. DODIG-2013-092, Release of Department of Defense Information to the
Media 12–14 (2013), available at http://www.dodig.mil/pubs/documents/DODIG-2013092.pdf (on ﬁle with the Columbia Law Review).
32. Such individuals may be receiving preferential treatment under current prepublication-review policies. See infra Part II.C.2.
33. Ian Shapira, CIA Memoirs Offer Revelations and Settle Scores Among Spies,
Wash. Post (June 4, 2012), http://articles.washingtonpost.com/2012-06-04/local/354611
65_1_cia-spymaster-john-kiriakou-publications-review-board (on ﬁle with the Columbia Law
Review) (“‘In many cases, [memoirs] are providing the only account there is, and people
read the memoirs to ﬂesh out a sparse public record.’”) (quoting Steven Aftergood,
Federation of American Scientists).
34. See The Constitution Project, The Report of the Constitution Project’s Task
Force on Detainee Treatment passim (2013), available at http://detaineetaskforce.org/
pdf/Full-Report.pdf (on ﬁle with the Columbia Law Review) (providing “comprehensive
record of detainee treatment across multiple administrations and multiple geographic
theatres” based on “public records and interviews with more than 100 people, including
former detainees, military and intelligence officers, interrogators and policymakers,”
based in part on accounts of detainee mistreatment written by former IC employees).
35. See, e.g., Mark Mansﬁeld, Reﬂections on Service: A Conversation with Former
CIA Director Michael Hayden, Stud. Intelligence, June 2010, at 63, 65–66 (“[W]hatever
you may think of this, it worked and we did indeed get life-saving intelligence out of it.”);
John Rizzo, CIA’s Enhanced Interrogation “Necessary and Effective,” Frontline (Sept. 13,
2011, 2:24 PM), http://www.pbs.org/wgbh/pages/frontline/iraq-war-on-terror/the-interro
gator/john-rizzo-cias-enhanced-interrogation-necessary-and-effective/ (on ﬁle with the
Columbia Law Review) (quoting former acting General Counsel of CIA in describing
enhanced interrogation techniques as “necessary and effective”).
424
COLUMBIA LAW REVIEW
[Vol. 115:417
claiming they produced “phenomenal” results.36 Such claims are difficult
to evaluate, as the details of individual cases and any intelligence derived
from them are, of course, classiﬁed. The ﬁrsthand experiences of
interrogators thus add a unique perspective to this debate.37 The CIA
itself appears to be wary of such ﬁrsthand accounts; it allegedly told a
prospective author undergoing prepublication review, “We will not allow
you to take the reader into the interrogation room. We will not allow you
to make the prisoner a human being. To the extent that we can, we will
take out anything that gives him a personality.”38 The availability of such
alternative, credible accounts can act as a check against excessive
manipulation of the national-security narrative by the government.
3. Insider Publications May Assist Future FOIA Requests. — Publications
by insiders may serve another tangential but valuable purpose—alerting
those requesting information through the Freedom of Information Act39
(FOIA) to the existence of documents that may be of interest to them. In
what has been called the problem of “prerequisite knowledge,” a valid
FOIA request for information requires a reasonably speciﬁc description
of the document sought, but if the requestor is not even aware of the
existence of the document, it is impossible to describe it.40 Unclassiﬁed
works by insiders may alert the general public to at least the existence, if
not details, of contentious programs or interpretations. These publications may thereby “make shallow” otherwise “deep secrets.”41 Insider
accounts are thus an important source of information that is difficult for
the public to obtain elsewhere. As described in the following section, the
constitutionality of prepublication review appears to be beyond question;
36. Chris McGreal, Dick Cheney Defends Use of Torture on Al-Qaida Leaders,
Guardian (Sept. 9, 2011, 1:17 PM), http://www.theguardian.com/world/2011/sep/09/
dick-cheney-defends-torture-al-qaida (on ﬁle with the Columbia Law Review).
37. See, e.g., Dina Temple-Raston, In ‘The Black Banners,’ Ali Soufan Takes Readers
Inside the Interrogation Room, Wash. Post (Oct. 28, 2011), http://www.washingtonpost.
com/entertainment/books/in-the-black-banners-ali-soufan-takes-readers-inside-the-interro
gation-room/2011/10/18/gIQAePOCQM_story.html (on ﬁle with the Columbia Law
Review) (“Soufan is not a journalist. The conversations he re-creates in the early part of the
book sound somewhat wooden and forced. But he redeems himself with detailed
descriptions of what unfolded behind the closed doors of the world’s interrogation
rooms.”).
38. Laura Miller, Censored by the CIA, Salon (Aug. 30, 2011, 8:31 AM), http://www.
salon.com/2011/08/31/censored_by_cia/ (on ﬁle with the Columbia Law Review).
39. 5 U.S.C. § 552 (2012) (mandating full or partial disclosure of unreleased information and documents controlled by government unless one of nine exemptions is met).
40. See Seth F. Kreimer, The Freedom of Information Act and the Ecology of
Transparency, 10 U. Pa. J. Const. L. 1011, 1025–27 (2008) (explaining unique challenge
presented by FOIA requirement of reasonable speciﬁcity in describing information
requested regarding classiﬁed national-security programs).
41. See David E. Pozen, Deep Secrecy, 62 Stan. L. Rev. 257, 274 (2010) (deﬁning
“deep secret” as one where “small group of similarly situated officials conceals its
existence from the public and from other officials, such that the outsiders’ ignorance
precludes them from learning about, checking, or inﬂuencing the keepers’ use of the
information”).
2015]
PREPUBLICATION REVIEW
425
it is thus particularly important that the review policies of IC agencies be
calibrated to achieve the maximum public disclosure possible without an
adverse impact on national security.
B.
The Constitutionality of Prepublication Review
There is no explicit statutory authority for prepublication review;
authority has been implied from section 3024 of the National Security
Act of 1947. 42 That text reads simply: “The Director of National
Intelligence shall protect intelligence sources and methods from
unauthorized disclosure.” 43 The concept originated with a set of
nondisclosure agreements that the CIA required its employees to sign as
a condition of employment.44 These agreements prevented employees
from seeking publication of their manuscripts without written consent
from the agency. Despite being an act of prior restraint,45 such agreements were upheld by the Supreme Court.46
1. The First Amendment and the Strong Presumption Against Prior
Restraint. — The First Amendment of the Constitution provides that
“Congress shall make no law . . . abridging the freedom of speech, or of
the press; of the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.”47 While some proponents of strong First Amendment protections have insisted that the
phrase “shall make no law” is unambiguous and means neither Congress
nor the courts can restrict speech,48 the Court has consistently held that
restraints on free expression may be “permitted for appropriate reasons.”49 Despite these exceptions, there is a special presumption against
42. 50 U.S.C. §§ 401–442 (2012).
43. Id. § 403(i)(1). By way of clariﬁcation, the reorganization of the IC in 2004
shifted this mandate, which was originally assigned to the CIA Director, to the Director of
National Intelligence. See Intelligence Reform and Terrorism Prevention Act of 2004,
Pub. L. No. 108-458, sec. 1011, § 102A(i), 118 Stat. 3638, 3651. Because some of the court
decisions that follow predate the reorganization, they make repeated references to the CIA
Director’s mandate under the act, not the ODNI Director’s.
44. See infra note 56–57 and accompanying text (describing requirements imposed
on CIA employee).
45. See, e.g., Alexander v. United States, 509 U.S. 544, 553 (1993) (“The doctrine of
prior restraint originated in the common law of England, where prior restraints of the
press were not permitted, but punishment after publication was. This very limited
application of the principle of freedom of speech was held inconsistent with our First
Amendment . . . .”).
46. Snepp v. United States, 444 U.S. 507, 507–08 (1980) (per curiam) (enforcing
former CIA agent’s employment agreement with Agency and imposing constructive trust
on book proﬁts). For academic works criticizing the decision, see infra note 81.
47. U.S. Const. amend. I.
48. See, e.g., Hugo L. Black, The Bill of Rights, 35 N.Y.U. L. Rev. 865, 874, 879
(1960) (“The phrase ‘Congress shall make no law’ is composed of plain words, easily
understood.”).
49. Elrod v. Burns, 427 U.S. 347, 360 (1976).
426
COLUMBIA LAW REVIEW
[Vol. 115:417
the practice of prior restraint, as articulated most famously in New York
Times Co. v. United States.50
Where the speech involves criticism of the government by its
employees, the Court has held that “the interests of the [employee], as a
citizen, in commenting upon matters of public concern” must be balanced against the interest of “the State as an employer, in promoting the
efficiency of the public services it performs through its employees.”51
Comments by government employees on matters of public concern cannot be restricted just because they are critical in tone.52 Courts have
placed particular emphasis on the phrase “matters of public concern”:
Where the employee’s expression “cannot be fairly considered as relating
to any matter of political, social, or other concern to the community . . .
officials should enjoy wide latitude in managing their offices, without
intrusive oversight by the judiciary in the name of the First
Amendment.”53 The issue is thus the balance between the public value of
the speech and the government’s interest in restricting it. In light of the
beneﬁts of insider publications articulated above,54 such works should
qualify as speech of public concern; the question, then, is the state’s
interest in restricting the employee’s speech.
2. The First Look: Marchetti. — The ﬁrst case to consider the
constitutionality of mandating prepublication review in a nondisclosure
agreement was United States v. Marchetti. 55 Victor Marchetti was a
fourteen-year employee of the CIA who signed a secrecy agreement56
both upon joining and resigning from the agency.57 After resigning, he
published a novel and several articles based on his experiences as an
agent, and the government sought an injunction against further publication.58 The Fourth Circuit upheld a district-court injunction requiring
Marchetti to submit, at least thirty days in advance, any writing related to
his experiences in the CIA.59 Importantly, the court limited its scope to
50. 403 U.S. 713, 714 (1971) (per curiam) (“‘Any system of prior restraints of
expression comes to this Court bearing a heavy presumption against its constitutional
validity.’” (quoting Bantam Books, Inc. v. Sullivan, 372 U.S. 58, 70 (1963))). This case is
also known as The Pentagon Papers Case.
51. Pickering v. Bd. of Educ., 391 U.S. 563, 568 (1968).
52. Id. at 570.
53. Connick v. Myers, 461 U.S. 138, 146 (1983).
54. See supra Part I.A (articulating beneﬁts of publications by former employees of
IC).
55. 466 F.2d 1309 (4th Cir.), cert. denied, 409 U.S. 1063 (1972).
56. Id. at 1312 (describing agreement through which Marchetti agreed “not to
divulge in any way any classiﬁed information, intelligence, or knowledge, except in the
performance of his official duties, unless speciﬁcally authorized in writing by the Director
or his authorized representative”).
57. Id.
58. Id. at 1313.
59. Id. at 1311.
2015]
PREPUBLICATION REVIEW
427
classiﬁed information.60 The Fourth Circuit recognized both the burden
on Marchetti and the executive’s right to secrecy61 and held that, given
the National Security Act of 1947’s requirement that the CIA director
protect intelligence sources and methods,62 “a system of prior restraint
against disclosure by employees and former employees of classiﬁed
information obtained during the course of employment” was a
reasonable means to protect such secrets.63
Applying this standard, the Fourth Circuit approved the secrecy
agreement signed by Marchetti upon joining the agency but declined to
enforce the agreement signed upon resignation, as it purported to
restrict his ability to disclose unclassiﬁed information.64 Additionally, the
court stated that if classiﬁed information were found “in the public
domain,” then Marchetti should have as much right as anyone else to
republish it. 65 Otherwise, the court found the requirement for
prepublication review valid, with a few caveats that remain important
today. First, recognizing prior restraint’s heavy burden on prospective
authors, the court held that the CIA “must act promptly to approve or
disapprove any material” submitted by authors, suggesting a maximum
response time of thirty days.66 Second, the court held that to sustain such
prior restraint, any author disagreeing with CIA prepublication decisions
should be entitled to judicial review,67 though it stated that such review
would not extend to the original decision to classify.68
3. The Supreme Court Weighs In: Snepp v. United States. — Several
years after the Marchetti decision, the Supreme Court weighed in on the
constitutionality of prepublication review in Snepp v. United States.69 Frank
60. Id. at 1318.
61. See id. at 1315 (“Citizens have the right to criticize the conduct of our foreign
affairs, but the Government also has the right and the duty to strive for internal secrecy
about the conduct of governmental affairs in areas in which disclosure may reasonably be
thought to be inconsistent with the national interest.”).
62. Id. at 1316.
63. See id. at 1317 (noting ex post criminal sanctions would be insufficient safeguard
in light of potential harm from disclosure).
64. Id.
65. Id. at 1318. The D.C. Circuit similarly held that the government may not censor
material already in the public domain, as it “has no legitimate interest in censoring
unclassiﬁed materials.” McGehee v. Casey, 718 F.2d 1137, 1141 (D.C. Cir. 1983).
66. Marchetti, 466 F.2d at 1317. This thirty-day period has remained inﬂuential in
guiding agency policy but has since been treated as more of a soft goal than a hard
requirement. See infra notes 156–158 and accompanying text.
67. Marchetti, 466 F.2d at 1317 (citing Freedman v. Maryland, 380 U.S. 51 (1965)
(striking down Maryland’s prior restraint regime for ﬁlms because it lacked express
guarantee of judicial review)).
68. Id. at 1318 (“The courts, of course, are ill-equipped to become sufficiently
steeped in foreign intelligence matters to serve effectively in the review of secrecy
classiﬁcations in that area.”).
69. 444 U.S. 507 (1980) (per curiam).
428
COLUMBIA LAW REVIEW
[Vol. 115:417
Snepp worked for the CIA for eight years70 and, despite having signed a
secrecy agreement upon employment, published a book71 based on his
experiences without seeking prepublication review.72 What distinguished
his agreement from Marchetti’s was a provision in which he agreed not to
publish any information—classiﬁed or not—relating to the agency, its
activities, or intelligence activities generally without prior approval.73
The Supreme Court held that Snepp’s employment with the CIA
involved “an extremely high degree of trust”74 and that such a special
trust relationship required that Snepp give the CIA “an opportunity to
determine whether the material he proposed to publish would compromise classiﬁed information or sources.”75 For the Court, it was irrelevant
whether or not the book actually contained classiﬁed information76—it
dispensed with Snepp’s First Amendment claim in a footnote. 77 The
Court then cited the ﬁndings of the district court, concluding that
“Snepp’s breach of his explicit obligation” had “irreparably harmed the
United States.”78 The Court placed a constructive trust on Snepp’s profits, calling it “the most appropriate remedy.”79
4. Post-Snepp: Expansion, Backlash, and Regularization. — Legal scholars had debated the use of nondisclosure agreements and prepublication
review even before the decision in Snepp.80 The decision drew immediate
criticism, almost all of which focused on the constitutionality of prior
70. Frank Snepp, Irreparable Harm: A Firsthand Account of How One Agent Took
on the CIA in an Epic Battle over Secrecy and Free Speech 60–61 (1999) [hereinafter
Snepp, Irreparable Harm].
71. Frank Snepp, Decent Interval: An Insider’s Account of Saigon’s Indecent End
Told by the CIA’s Chief Strategy Analyst in Vietnam (1977).
72. Snepp, 444 U.S. at 507.
73. See id. at 508 (“Thus, Snepp had pledged not to divulge classiﬁed information
and not to publish any information without prepublication clearance.”).
74. Id. at 510.
75. Id. at 511.
76. Id.
77. Id. at 509 n.3.
78. Id. at 513. The Court placed particular emphasis on the statement of Stansﬁeld
Turner, then the CIA Director, about the loss of valuable intelligence sources due to
Snepp’s book. Id. at 512–13. Ironically enough, Turner later became an outspoken critic
of the CIA’s prepublication-review procedures after his own memoir was put through the
process. See infra note 207 and accompanying text.
79. Snepp, 444 U.S. at 514–16.
80. See, e.g., James Peter Rau, Government Secrecy Agreements and the First
Amendment, 28 Am. U. L. Rev. 395, 396–97 (1979) (analyzing two First Amendment
doctrines, reasonableness standard and prior restraint, in context of prepublication
review); Joshua B. Bolten, Comment, Enforcing the CIA’s Secrecy Agreement Through
Postpublication Civil Action: United States v. Snepp, 32 Stan. L. Rev. 409, 410 (1980)
(arguing district court’s decision in Snepp was correct in terms of remedy); Caroline Heck,
Comment, National Security and the First Amendment: The CIA in the Marketplace of
Ideas, 14 Harv. C.R.-C.L. L. Rev. 655, 658 (1979) (comparing secrecy agreements with
nondisclosure agreements in private sector and concluding courts have given too much
weight to government assertions of national-security issues).
2015]
PREPUBLICATION REVIEW
429
restraint on current and former CIA employees.81 Some scholars and
practitioners urged Congress to act to counter the effects of the decision.82 Instead, the decision was seized upon and became the catalyst for
a major change in executive-branch policy. President Reagan relied on
the reasoning in Snepp to pass National Security Decision Directive 84
(NSDD-84), which expanded the requirement to sign a nondisclosure
agreement mandating prepublication review to all employees of the
executive branch. 83 In response to heavy criticism from Congress, 84
President Reagan suspended NSDD-84’s lifetime prepublication-review
requirement in February 1984.85 But this change “had little effect on
prepublication review requirements” because employees were still
required to sign other agreements before being granted access to sensitive compartmented information (SCI)86—information protected by a
higher level of classiﬁcation. 87 After years of public disagreement
81. See, e.g., Thomas M. Franck & James J. Eisen, Balancing National Security and
Free Speech, 14 N.Y.U. J. Int’l L. & Pol. 339, 339–43 (1982) (asserting intent to “answer
those questions nunc pro tunc that were not addressed in Snepp” and concluding there
should be no ﬁduciary duty to submit to prepublication review where information being
published is unclassiﬁed); Jonathan C. Medow, The First Amendment and the Secrecy
State: Snepp v. United States, 130 U. Pa. L. Rev. 775, 840 (1982) (arguing case was wrongly
decided based on case law concerning prior restraint and should be reconsidered at ﬁrst
opportunity); Diane F. Orentlicher, Comment, Snepp v. United States: The CIA Secrecy
Agreement and the First Amendment, 81 Colum. L. Rev. 662, 706 (1981) (arguing
decision goes against precedent in two particular areas: prior restraint and government
restrictions of public employees’ speech).
82. See, e.g., Comm. on Fed. Legislation, The Response to Snepp v. United States: A
Proposal for the 97th Congress, 36 Rec. Ass’n B. City N.Y. 299, 299 (1981) (urging
legislature to act to guarantee more robust First Amendment protections for employees of
national-security organizations).
83. President Ronald Reagan, National Security Decision Directive No. 84, Safeguarding National Security Information 1 (Mar. 11, 1983), available at http://www.reagan.
utexas.edu/archives/reference/Scanned%20NSDDS/NSDD84.pdf (on ﬁle with the
Columbia Law Review). The directive also emphasized an increase in the use of polygraph
technology. Id. at 2–3.
84. See Louis Fisher, Congressional–Executive Struggles over Information: Secrecy
Pledges, 42 Admin. L. Rev. 89, 92–94 (1990) (describing hearings held by Senate
Committee on Governmental Affairs over constitutionality of directive, during which
several senators expressed disapproval).
85. U.S. Gov’t Accountability Office, No. GAO/NSIAD-91-106FS, Information Security: Federal Agency Use of Nondisclosure Agreements 10 (1991), available at http://www.
gao.gov/assets/90/89057.pdf (on ﬁle with the Columbia Law Review).
86. U.S. Gov’t Accountability Office, No. GAO/T-NSIAD-88-44, Classiﬁed Information Nondisclosure Agreements: Statement of Louis J. Rodrigues, Associate Director,
National Security and International Affairs Division, Before the Subcommittee on Legislation and International Security, Committee on Government Operations, United States
House of Representatives 3 (1988), available at http://www.gao.gov/assets/110/102256
.pdf (on ﬁle with the Columbia Law Review).
87. See Office of the Dir. of Nat’l Intelligence, Intelligence Community Directive No.
703, Protection of Classiﬁed National Intelligence, Including Sensitive Compartmented
Information 2 (June 21, 2012) [hereinafter IC Directive 703], available at http://fas.org/
irp/dni/icd/icd-703.pdf (on ﬁle with the Columbia Law Review) (deﬁning sensitive
430
COLUMBIA LAW REVIEW
[Vol. 115:417
between Congress and the executive88—as well as several lawsuits by
executive-branch employees89—nondisclosure forms mandating prepublication review were still being utilized at the end of the Reagan
Administration.90 Thus, despite the initial public outcry over the proposal
to expand the applicability of prepublication review,91 such agreements
became the norm for federal employees with access to sensitive information and have remained so.92
This Note does not question the constitutionality of such forms.
Instead, it assesses the review process as actually practiced by IC agencies.
It compares the policies and practices of all IC agencies, but focuses
primarily on the CIA because it appears to be both the source of the
highest number of controversies surrounding prepublication review 93
and—perhaps as a result—the agency with the most information publicly
available about its review practices.
II. PREPUBLICATION REVIEW IN PRACTICE
Despite the apparent need for cohesive executive-branch policy, IC
agencies are granted wide discretion in formulating their own
prepublication-review practices. Part II explores the prepublicationcompartmented information as “subset of [classiﬁed information] concerning or derived
from intelligence sources, methods or analytical processes”).
88. See Fisher, supra note 84, at 92 (outlining disagreements between Congress and
executive branch over expanded use of mandatory nondisclosure forms).
89. See Am. Foreign Serv. Ass’n v. Garﬁnkel, 490 U.S. 153, 161 (1989) (per curiam)
(reversing lower court’s expansive ruling on power of executive and instructing it to
exercise restraint); Nat’l Fed’n of Fed. Emps. v. United States, 695 F. Supp. 1196, 1202–03
(D.D.C. 1988) (ﬁnding term “classiﬁable,” used in nondisclosure forms, to be
insufficiently narrow to survive constitutional scrutiny); Nat’l Fed’n of Fed. Emps. v.
United States, 688 F. Supp. 671, 676 (D.D.C. 1988) (addressing plaintiffs’ objections to
language contained in SF 189, SF 4193, and related forms).
90. Employees were still required to sign such nondisclosure agreements to gain
access to SCI. See supra note 86 and accompanying text (explaining NSDD-84 still
required employees to sign a nondisclosure agreement to access SCI).
91. Public outrage seems to have subsided rather quickly. See Donna A. Demac,
Hearts and Minds Revisited: The Information Policies of the Reagan Administration, in
Vincent Mosco & Janet Wasko, The Political Economy of Information 125, 129 (1988) (“If
the administration anticipated that a population accustomed to news as entertainment
and sensation would quickly grow bored with prepublication review . . . it was correct. The
storm over NSDD 84 gave way to the eerie quiet surrounding government moves to
[enforce its policies].”).
92. See, e.g., DIA Employment Conditions, supra note 4 (requiring employees to sign
agreement “not to disclose, in any fashion, classiﬁed information to unauthorized
persons,” which also authorizes “Agency pre-publication review of certain material prior to
disclosure during and after employment with the DIA”).
93. Although no single source has compiled comprehensive statistics on prepublication-review controversies among executive-branch agencies, an extensive review of relevant
materials revealed that the majority of disputes that reach the courts or are covered by
news outlets involve disagreement with the CIA. See infra Part II.C (discussing problematic
practices of prepublication review).
2015]
PREPUBLICATION REVIEW
431
review policies of IC agencies. Part II.A discusses the lack of executivebranch-wide guidance on review policies. Part II.B looks at the official
policies of IC agencies, with II.B.1 focusing on the CIA, II.B.2 on the
Department of Defense (DoD), and II.B.3 on other IC agencies. Part II.C
looks at the actual implementation of these policies and illuminates
potential issues. Part II.C.1 discusses the risk of opportunistic post hoc
classiﬁcation decisions. Part II.C.2 argues that the experiences of authors
suggest agencies practice selective enforcement of their review policies.
Finally, Part II.C.3 explores the risk of intimidation and delay tactics.
A.
The Lack of Executive-Branch-Wide Policies
There is no explicit statutory authority for prepublication review,94
and, more surprisingly, there is no executive-branch-wide policy governing prepublication review. Executive Order 13,526 (“Classiﬁed National
Security Information”) requires that a nondisclosure agreement be
signed before access to classiﬁed information is granted95 but otherwise
does little more than parrot the language of the National Security Act of
1947. 96 The regulation implementing Executive Order 13,526 does
require agencies to utilize a particular nondisclosure form: Standard
Form (SF) 312.97 SF 312 itself does not mandate prepublication review;98
it only requires that the signatory not disclose classiﬁed information to
unauthorized recipients.99 Thus, access to classiﬁed information does not
automatically trigger a prepublication-review obligation. In reality, however, most IC agencies have internal policies that, at the very least,
require current employees to submit any proposed works for prepublication review.100 Many agencies require lifetime prepublication review as a
94. See supra notes 42–46 and accompanying text (discussing basis for prepublication review).
95. Exec. Order No. 13,526, 3 C.F.R. 298, 314 (2010).
96. Compare id. at 723 (granting primary implementation authority to ISOO, but
providing ODNI with authority, after consultation with heads of affected agencies and
Director of ISOO, to issue guidance directives with respect to protection of intelligence
sources, methods, and activities), with National Security Act of 1947, 50 U.S.C. § 403(i)(1)
(2012) (“The Director of National Intelligence shall protect intelligence sources and
methods from unauthorized disclosure.”).
97. See 32 C.F.R. § 2001.80(d)(2) (2014) (describing requirements of SF 312).
Agencies may seek a waiver from ODNI to utilize a different form. Id.
§ 2001.80(d)(2)(viii).
98. See Info. Sec. Oversight Office, Classiﬁed Information Nondisclosure Agreement
(Standard Form 312) Brieﬁng Booklet 73 (Spring 2001 reprt.), available at http://www.
wrc.noaa.gov/wrso/forms/standard-form-312_booklet.pdf (on ﬁle with the Columbia Law
Review) (“There is no explicit or implicit prepublication review requirement in the SF
312 . . . .”).
99. See Standard Form No. 312, supra note 8 (requiring signatory to agree not to
divulge classiﬁed information unless signatory has “officially veriﬁed that the recipient has
been properly authorized by the United States Government to receive it” or has obtained
waiver).
100. See infra Part II.B (describing prepublication-review policies of IC agencies).
432
COLUMBIA LAW REVIEW
[Vol. 115:417
condition of employment.101 Other agencies have no such conditions,
but require employees to sign additional nondisclosure agreements—
usually as a precondition to accessing SCI—that do incur lifetime
prepublication-review obligations.102 Thus, even where an agency does
not categorically require prepublication review, employees may incur a
lifetime obligation.
For this reason, agencies have discretion in determining which types
of employees incur prepublication-review obligations, but their discretion does not end there: Because there is no executive-branch-wide
policy outlining what prepublication review should entail, individual
agencies determine the scope of prepublication review.103 Two offices—
the Information Security Oversight Office (ISOO) and the Office of the
Director of National Intelligence (ODNI)—seemingly possess the authority to promulgate guidance but, as discussed next, have not done so.
ISOO—established in 1978 as a component of the National Archives
and Records Administration—is the primary body that oversees the
classiﬁcation system.104 Its mission is to “support the President by ensuring that the Government protects and provides proper access to information to advance the national and public interest” and to “lead efforts
to standardize and assess the management of classiﬁed and controlled
unclassiﬁed information through oversight, policy development, guidance, education, and reporting.”105 None of its guidance on implementing Executive Order 13,526 addresses prepublication review.106
101. The CIA, DIA, and NSA do so. See CIA, Agency Prepublication Review of Certain
Material Prepared for Public Dissemination (May 30, 2007) [hereinafter 2007 CIA Policy],
available at https://www.fas.org/irp/cia/prb2007.pdf (on ﬁle with the Columbia Law
Review) (establishing procedures for CIA prepublication review); DIA Employment
Conditions, supra note 4 (requiring employees to sign agreement authorizing
prepublication review); NSA, Statement of Conditions of Employment (June 2009),
available at https://www.nsa.gov/careers/_ﬁles/P2771.pdf (on ﬁle with the Columbia Law
Review) (same).
102. See Form 4414, Sensitive Compartmented Information Nondisclosure Agreement
(Dec. 2013), http://www.ncix.gov/SEA/docs/FORM_4414_Rev_12_2013.pdf (on ﬁle with
the Columbia Law Review) (requiring, in exchange for access to SCI, prepublication review
of “any writing or other preparation in any form, including a work of ﬁction, that contains
or purports to contain any SCI or description of activities that produce or relate to SCI”).
ODNI is currently in the process of updating IC Form 4414. Info. Sec. Oversight Office,
ISOO Notice 2013-05: Revision of the Standard Form (SF) 312, “Classiﬁed Information
Nondisclosure Agreement” 2 (Aug. 19, 2013), available at http://www.archives.gov/isoo/
notices/notice-2013-05.pdf (on ﬁle with the Columbia Law Review).
103. See infra Part II.B (discussing disparate agency policies).
104. Elizabeth Goitein & David M. Shapiro, Brennan Ctr. for Justice, Reducing
Overclassiﬁcation Through Accountability 19 (2011), available at http://www.brennancen
ter.org/sites/default/ﬁles/legacy/Justice/LNS/Brennan_Overclassiﬁcation_Final.pdf (on
ﬁle with the Columbia Law Review).
105. 2013 ISOO Report, supra note 19, at i.
106. See 32 C.F.R § 2001 (2014) (laying out review standards for agency classiﬁcation,
declassiﬁcation, and safeguarding of national-security information).
2015]
PREPUBLICATION REVIEW
433
Following the recommendations of the 9/11 Commission Report,107
Congress created ODNI to unify and act as the head of the IC.108 Many of
the responsibilities previously delegated to the Director of Central
Intelligence were transferred to ODNI, including the mandate “to protect intelligence sources and methods from unauthorized disclosure.”109
None of the publicly available implementing directives of ODNI provide
executive-branch-wide guidance on prepublication review. Beyond requiring the heads of the IC elements to ensure that all employees have
signed a nondisclosure agreement and are “advised of [their] legal and
administrative obligations and the ramiﬁcations of a failure to meet those
obligations,”110 an ODNI directive addressing the protection of nationalsecurity information says nothing about prepublication review. Another
directive on the protection of classiﬁed information is similarly silent.111
The only policy document addressing prepublication review is an ODNI
Instruction that applies only to employees who work directly for ODNI or
are detailed there.112 The lack of executive-branch-wide guidance leaves
individual agencies responsible for developing their own policies. The
following sections will analyze, to the extent possible, the review policies
of IC agencies.
B.
Official IC Prepublication Policies
1. The CIA. — The Publications Review Board (PRB) is responsible
for the CIA’s prepublication review of current and former employees’
proposed publications.113 Unlike those of some agencies, the CIA’s review
107. Nat’l Comm’n on Terrorist Attacks upon the U.S., The 9/11 Commission Report
411 (2004), available at www.9-11commission.gov/report/911Report.pdf (on ﬁle with the
Columbia Law Review).
108. Intelligence Reform and Terrorist Prevention Act of 2004, Pub. L. No. 108-458,
sec. 1011, § 102A(i), 118 Stat. 3638, 3651.
109. 50 U.S.C. § 403(i)(1) (2012); see also supra note 43 and accompanying text (explaining transfer of authority).
110. Office of the Dir. of Nat’l Intelligence, Intelligence Community Directive No.
700, Protection of National Intelligence 3 (June 7, 2012), available at http://www.dni.gov/
ﬁles/documents/ICD/ICD_700.pdf (on ﬁle with the Columbia Law Review).
111. See IC Directive 703, supra note 87 (addressing classiﬁcation issues speciﬁc to SCI
but not mentioning prepublication review).
112. See Office of the Dir. of Nat’l Intelligence, Instruction 80.04: ODNI PrePublication Review of Information to Be Publicly Released 1 (Apr. 8, 2014) [hereinafter
ODNI, Instruction 80.04], available at http://www.dni.gov/ﬁles/documents/CIO/Instr.%
2080.04%20%20Pre-Publication%20Review%20of%20Information%20to%20be%20Publicl
y%20Relea.pdf (on ﬁle with the Columbia Law Review) (addressing scope of applicability).
113. See John Hollister Hedley, Reviewing the Work of CIA Authors: Secrets, Free
Speech, and Fig Leaves, Stud. Intelligence, Spring 1998, at 75, 75, available at https://
www.cia.gov/library/center-for-the-study-of-intelligence/kent-csi/vol41no5/pdf/v41i5a04
p.pdf (on ﬁle with the Columbia Law Review) (explaining role of PRB in prepublication
review for CIA).
434
COLUMBIA LAW REVIEW
[Vol. 115:417
policies are not publicly available.114 The most recent PRB regulation
accessible is a heavily redacted version dated May 30, 2007, and marked
as approved for release in January 2008.115 The regulations were updated
in 2011116 but have not been publicly released. Despite the lack of official
policy documents, the CIA’s frequent involvement in litigation over
prepublication-review decisions and the apparent willingness of former
PRB officials to describe the process means there is a surprising abundance of documents that may provide more insight into actual review
policies than do the official policy documents of other agencies.
The PRB was created in 1976 but has been around in its current
form only since 2007, when the CIA approved the creation of a full-time,
fully staffed board.117 This was in response to a large increase in the number of submissions—from 1980 to 2003, the Board reviewed between 200
and 400 manuscripts per year, while in 2010 the Board received over
1,800 and was anticipating more than 2,500 submissions for 2011.118 The
Board consists of a Chair and an Executive Secretary, designated by and
reporting directly to the Chief of Information Management Services.119
The rest of the Board is composed of senior representatives from each of
the directorates, including those offices under the Office of the
Director.120 The Office of General Counsel (OGC) provides a nonvoting
legal advisor.121 Any decision to “deny-in-full” a manuscript for publication requires a majority vote of the Board,122 which seems to suggest that
no single directorate has disproportionate inﬂuence over the Board’s
decisions. The decision of the PRB to deny a manuscript in full or in part
may be appealed. The author may submit additional materials and
request reconsideration by the Board; if that request is denied, the
author may appeal to the Associate Deputy Director of the CIA
114. For example, the NSA’s policies on prepublication review are available on its
website. NSA, NSA/CSS Policy 1-30: Review of NSA/CSS Information Intended for Public
Release (May 10, 2013) [hereinafter NSA Policy], available at http://www.nsa.gov/public_
info/_ﬁles/nsacss_policies/Policy_1-30.pdf (on ﬁle with the Columbia Law Review).
115. 2007 CIA Policy, supra note 101.
116. See CIA Prepublication Review in the Information Age, Stud. Intelligence, Sept.
2011, at 9, 9–11, available at http://www2.gwu.edu/~nsarchiv/NSAEBB/NSAEBB431/doc
s/intell_ebb_018.PDF (on ﬁle with the Columbia Law Review) (attempting to correct misconception of regulations released in June 2011 as “strictest version yet”).
117. Id.
118. Id. at 11. The author speculates that this sudden rise is due to a conﬂuence of
factors, including the proliferation of new publishing platforms, the introduction of new
agency policies encouraging employees to broaden their perspectives through outside
contacts, and the public’s increasing interest in intelligence. Id.
119. 2007 CIA Policy, supra note 101, at sec. 2(c)(1).
120. Id. For an organizational chart of CIA leadership, see CIA Organization Chart,
CIA, https://www.cia.gov/about-cia/leadership/ciaorgchart.jpg/image.jpg (on ﬁle with
the Columbia Law Review) (last updated Feb. 18, 2014, 12:42 PM).
121. 2007 CIA Policy, supra note 101, at sec. 2(c)(1).
122. CIA Prepublication Review in the Information Age, supra note 116, at 15–16.
2015]
PREPUBLICATION REVIEW
435
(ADD/CIA) within thirty days.123 The decision of the ADD/CIA, who
considers a recommendation from the OGC, is ﬁnal.124
The CIA requires submission of all “intelligence-related materials
intended for publication or public dissemination,” explicitly excluding
from its purview material “unrelated to intelligence, foreign relations, or
CIA employment or contract matters.”125 It distinguishes between official
and unofficial publications. Unofficial publications refer to works written
by an employee who has signed “a CIA secrecy agreement” and who has
prepared the work as a private individual and not in any official capacity.126 Official publications are deﬁned as works intended to be unclassiﬁed and prepared as part of the employee’s official duties.127
In addition to the distinction between official and unofficial works,
the CIA takes a bifurcated approach to review by treating current and
former employees differently. Until July 2005, the PRB only reviewed
publications by former employees; current employees had their works
reviewed by their immediate supervisors.128 The PRB now reviews works
by both but applies different standards. Publications by former employees are reviewed “solely to determine whether it contains any classiﬁed
information.”129 In contrast, the PRB is authorized to consider additional
factors when reviewing the work of current employees. It may deny
publication of information that could “reasonably be expected to impair
the author’s performance of his or her job duties; interfere with the
authorized functions of the CIA, or; have an adverse effect on the foreign
relations or security of the U.S.”130 Furthermore, the employee must
obtain her immediate supervisor’s concurrence (or, in the case of
contractors, the contracting officer’s concurrence) that the material is
appropriate for publication.131
The CIA clearly states that it considers additional factors in reviewing a current employee’s proposed publication.132 But it seems to insist,
123. 2007 CIA Policy, supra note 101, at sec. 2(h)(1).
124. Id.
125. Id. at sec. 2(b)(1), (3).
126. Id. at sec. 2(b)(6). Accordingly, both current and former employees can author an
unofficial publication.
127. Id. at sec. 2(b)(7). Accordingly, only current employees can author an official
publication.
128. CIA Prepublication Review in the Information Age, supra note 116, at 10. But the
PRB could take on review of a current employee’s manuscript at the request of the
employee’s supervisors. Hedley, supra note 113, at 79.
129. 2007 CIA Policy, supra note 101, at sec. 2 (f)(2) (emphasis added).
130. Id. at sec. 2(g)(2).
131. Id. at sec. 2(g)(4)(a).
132. A recent example illuminates the importance of this distinction. Bridget Nolan
was a graduate fellow for the CIA and sought to publish a dissertation based on her
experiences there. She submitted a draft proposal to the PRB, and, despite her insistence
that it contained no classiﬁed information, the draft proposal was rejected because it was
deemed “inappropriate” for a current employee. Her dissertation was approved with
436
COLUMBIA LAW REVIEW
[Vol. 115:417
both in PRB policy documents and by implication in individual cases,
that the works of former employees are reviewed solely for classiﬁed
information. 133 The opinions of former employees who have gone
through the prepublication-review process, as well as those of PRB officials themselves, suggest otherwise. They indicate the process is not based
on such a simple, bright-line search for classiﬁed information but rather
on an amalgam of varying factors that leads to inconsistent treatment
and leaves open the possibility for abuse.134
2. Department of Defense Agencies. — Eight of the sixteen IC agencies
fall under DoD, making it the largest presence in the IC.135 DoD issues
prepublication-review guidance applicable to all of these agencies. 136
Similar to CIA policy, it distinguishes between official and unofficial
publications, requiring review of official publications in all cases and of
unofficial publications when criteria outlined in a separate implementing
instruction are met.137 It also distinguishes between current and former
employees, 138 but the scope of review does not seem to differ—the
directive simply states, “[P]ublic release of official DoD information is
limited only as necessary to safeguard information requiring protection
in the interest of national security or other legitimate government inter“miniscule” changes after she resigned. Susan Snyder, Covering the Undercovers, Phila.
Inquirer (Aug. 20, 2013), http://articles.philly.com/2013-08-20/news/41425761_1_cia-nat
ional-counterterrorism-center-sociology (on ﬁle with the Columbia Law Review).
133. See id. (describing employee’s experiences and quoting CIA spokesman as saying
current and former employees are subject to different rules); see also 2007 CIA Policy,
supra note 101, at sec. 2(f)(2) (explaining work by former employee is to be reviewed
“solely to determine whether it contains any classiﬁed information”).
134. See infra Part II.C.2 (discussing authors’ experiences with prepublication-review
process).
135. See Members of the IC, Office of the Dir. of Nat’l Intelligence, http://www.dni.
gov/index.php/intelligence-community/members-of-the-ic (on ﬁle with the Columbia Law
Review) (last visited Oct. 18, 2014) (outlining IC membership).
136. Dep’t of Def., DoD Directive No. 5230.09: Clearance of DoD Information for
Public Release 1 (Aug. 22, 2008) [hereinafter DoD Directive], available at http://www.dtic
.mil/whs/directives/corres/pdf/523009p.pdf (on ﬁle with the Columbia Law Review). In
addition to the intelligence agencies of each military branch, the NSA, the National
Geospatial-Intelligence Agency (NGA), the National Reconnaissance Office (NRO), and
the DIA are components of DoD. See Dep’t of Def., Organization of the Department of
Defense (DoD) (Mar. 2012), available at http://odam.defense.gov/Portals/43/Document
s/Functions/Organizational%20Portfolios/Organizations%20and%20Functions%20Guide
book/DoD_Organization_March_2012.pdf (on ﬁle with the Columbia Law Review) (providing organizational chart of DoD).
137. DoD Directive, supra note 136, at 2. For criteria requiring submission of
unofficial works, see Dep’t of Def., DoD Instruction Number 5230.29: Security and Policy
Review of DoD Information for Public Release 6 (Aug. 13, 2014) [hereinafter DoD
Instruction], available at http://www.dtic.mil/whs/directives/corres/pdf/523029p.pdf
(on ﬁle with the Columbia Law Review) (requiring submission where information, inter
alia, “[i]s or has the potential to become an item of national or international interest” or
“[a]ffects national security policy, foreign relations, or ongoing negotiations”).
138. DoD Directive, supra note 136, at 2 (requiring “[r]etired personnel, former DoD
employees, and non-active duty members of the Reserve Components” to submit works).
2015]
PREPUBLICATION REVIEW
437
est.”139 The Washington Headquarters Service is responsible for the system of review, 140 and within it, the Office of Security Review (OSR)
conducts the actual review.141
In addition to these generally applicable directives and instructions,
each DoD agency may promulgate its own policy. The NSA’s
prepublication-review policy 142 —which is publicly available—requires
both current and former employees to submit their materials to a
“Prepublication Review Authority” (PRA).143 The distinctions between
current and former employees and between official and unofficial works
seem to mirror those of the CIA—official publications are checked for
conformity to “NSA/CSS corporate messaging standards.”144 But unlike
CIA employees, current NSA employees must obtain initial review before
submission to the PRA—a Staff Security Officer must assess the potential
operational-security threat from disclosing one’s affiliation with NSA, and
a local Classiﬁcation Advisory Officer (CAO) makes an initial determination on whether the manuscript contains any classiﬁed information.145
The policy says the PRA “will issue, as practicable, a ﬁnal determination . . . within 25 business days of receipt”146 but says nothing about the
timeline for initial review by CAOs, creating the potential for delay. An
employee may appeal the PRA’s decision to the Associate Director for
Public Relations, who may consult with the OGC and/or the “information owners” before making its ﬁnal decision, which cannot be further
appealed.147
The Defense Intelligence Agency’s (DIA) basic approach is the same
as those of the CIA and DoD at large.148 The DIA does not publish its
139. Id. at 2. A DoD FAQ on prepublication review states review is necessary “to ensure
information damaging to the national security is not inadvertently disclosed.” Dep’t of
Def., Frequently Asked Questions for Department of Defense Security and Policy Reviews
1, http://www.dtic.mil/whs/esd/osr/docs/OUSD%28I%29Pre-PubPamphletFAQs%2820
12%29.pdf (on ﬁle with the Columbia Law Review) (last visited Oct. 31, 2014).
140. DoD Instruction, supra note 137, at 4.
141. See Defense Office of Prepublication and Security Review, Wash. Headquarters
Servs., http://www.dtic.mil/whs/esd/osr/ (on ﬁle with the Columbia Law Review) (last
visited Oct. 18, 2014) (“The Office of Security Review conducts the security and policy
review for clearance of official Department of Defense (DoD) information proposed for
official public release by the DoD and its employees (military and civilian).”).
142. The NSA is a component of DoD. See Exec. Order No. 12,333, 46 Fed. Reg.
59,941, 59,946–47 (Dec. 4, 1981) (requiring Secretary of Defense to “[d]irect, operate,
control and provide ﬁscal management for the National Security Agency”).
143. NSA Policy, supra note 114, at 7 (establishing responsibilities for both current
and former affiliates of NSA).
144. Id. at 4.
145. Id.
146. Id. at 5.
147. Id. at 6.
148. Both former and current employees are required to submit manuscripts, and
official and unofficial publications are treated differently. See Def. Intelligence Agency,
Instruction: Prepublication Review of Information Prepared for Public Release 7 (Aug. 18,
438
COLUMBIA LAW REVIEW
[Vol. 115:417
prepublication-review policies, but a 2006 version was made available
through a FOIA request. The organizations responsible for each step of
review are withheld under applicable FOIA exemptions, but the general
process can be discerned from the document. To ensure adequate
review, the primary office responsible for review coordinates with DoD’s
OSR, other IC elements with a stake in the information, and DIA elements with subject-matter expertise.149
3. Other Agencies. — Employees of ODNI150—which oversees the IC—
are also subject to mandatory review.151 With the exception of the State
Department, the remaining IC agencies do not require prepublication
review as a condition of employment—as discussed next, such an obligation arises only upon signing a nondisclosure agreement mandating it.
The State Department’s Foreign Affairs Manual requires both current and former employees to submit their materials to either the
“Bureau of Public Affairs [or the] Chief of Mission,” depending upon the
employee’s location,152 if such materials touch upon matters of “official
concern,” which appears to be deﬁned quite broadly153 but has been
upheld by courts.154 The review of works by former employees is described
as “limited” and is conducted “in accordance with applicable postemployment regulations and agreements.”155 Interestingly, the regulation
states that the duration of review is “not to exceed thirty days,”156 creating
a regulatory timeline that is more categorical than the regulations of
other agencies157 and that, if followed, may make it less likely that delay
can be used to dissuade publication.158
2006) [hereinafter DIA Policy], available at http://www.dia.mil/Portals/27/Documents/
FOIA/5%20USC%20%C2%A7%20552(A)(2)(C)%20Records/PREPUBLICATION%20REV
IEW%20OF%20INFORMATION%20PREPARED%20FOR%20PUBLIC%20REL.pdf (on ﬁle
with the Columbia Law Review) (outlining policy for nonofficial publications).
149. See id. at 3–4 (outlining responsibilities of redacted office).
150. See supra notes 107–109 and accompanying text (explaining ODNI’s creation).
151. See ODNI, Instruction 80.04, supra note 112, at 1 (deﬁning scope of applicability).
152. Dep’t of State, 3 FAM 4170: Official Clearance of Speaking, Writing, and
Teaching 3 (June 9, 2009) [hereinafter State Policy], available at http://www.state.gov/do
cuments/organization/85123.pdf (on ﬁle with the Columbia Law Review).
153. See id. (“Materials are on matters of official concern if they relate to any policy,
program, or operation of the employee’s agency or to current U.S. foreign policies, or
reasonably may be expected to affect the foreign relations of the United States.”
(emphasis added)).
154. See, e.g., Weaver v. U.S. Info. Agency, 87 F.3d 1429, 1431–32 (D.C. Cir. 1996)
(noting and accepting broad deﬁnition of “official concern”).
155. State Policy, supra note 152, at 4.
156. Id. at 5.
157. Most other agencies add qualifying language to make the deadline less absolute.
See, e.g., 2007 CIA Policy, supra note 101, at sec. 2(d)(4) (“Lengthy or complex
submissions may require a longer period of time for review . . . .”); DoD Instruction, supra
note 137, at 8 (“More time may be needed if . . . the material is complex or requires review
2015]
PREPUBLICATION REVIEW
439
The FBI and the Drug Enforcement Agency (DEA) are, as Justice
Department entities, governed by the same policy. Under it, prepublication review is required only to the extent it is expressly provided for in
nondisclosure agreements; the “nature and the extent of material” that
must be submitted for review is also determined by such agreements.159
The FBI has a Prepublication Review Office, which reviews the submitted
work and determines whether “further review is required and by
whom.”160
Other agencies similarly require prepublication review only upon
signing nondisclosure forms requiring it. The Department of Homeland
Security (DHS) has issued one directive addressing prepublication
review, but it merely says that an employee must sign a particular form to
gain access to SCI161 and that doing so incurs a lifetime obligation.162 It is
thus unclear whether DHS requires even current employees to submit
proposed publications for prepublication review—though it is likely, as
most other IC agencies, including the CIA and DoD, do so. 163 The
Department of Energy, for example, provides little insight into its
prepublication-review policies but does explicitly require current employees with access to Restricted Data (RD) or Formerly Restricted Data
(FRD) to submit to review. 164 The Treasury Department’s Office of
Terrorism and Financial Intelligence (TFI), however, lacks even that—
the most relevant provision of the Code of Federal Regulations, part 2
(“National Security Information”) of title 31 (“Money and Finance:
Treasury”), is silent on the topic of prepublication review165—and none
by agencies outside of the DoD.”); NSA Policy, supra note 114, at 6 (requiring, “as
practicable,” return of decision within twenty-ﬁve days).
158. See infra Part II.C.2–3 (alleging undue delay as agency tactic to prevent
publication).
159. 28 C.F.R. § 17.18(c)–(d) (2014).
160. Prepublication Review Office, FBI, http://www.fbi.gov/foia/prepublication-revi
ew-office (on ﬁle with the Columbia Law Review) (last visited Oct. 18, 2014).
161. See Dep’t of Homeland Sec., Management Directive No. 11043: Sensitive Compartmented Information Program Management 8 (Sept. 17, 2004), available at https://
www.dhs.gov/xlibrary/assets/foia/mgmt_directive_11043_sensitive_compartmented_infor
mation_program_management.pdf (on ﬁle with the Columbia Law Review) (“As a
condition of access to SCI, individuals must sign a DCI-authorized SCI Nondisclosure
Agreement (NdA) (Form 4414), which includes a provision for prepublication review.”).
162. See id. at 9 (“Persons who are currently, or were previously, employed by DHS and
indoctrinated for SCI access, will submit proposed articles and publications for
prepublication review.” (emphasis added)).
163. See supra Part II.B.1–2 (discussing agency requirements for current employees of
CIA and DoD).
164. 10 C.F.R. § 1045.44 (2014) (“Any person with authorized access to RD or FRD who
generates a document intended for public release in an RD or FRD subject area shall
ensure that it is reviewed for classiﬁcation . . . .” (emphasis added)).
165. See 31 C.F.R. pt. 2 (2014) (providing guidance for mandatory declassiﬁcation
review and granting historical researchers and former officials access to classiﬁed
information).
440
COLUMBIA LAW REVIEW
[Vol. 115:417
of the orders or directives issued by the department cover the topic of
prepublication review (or even nondisclosure agreements). 166 The
introduction to a memoir written by a former Treasury Department special agent suggests, however, that some form of prepublication-review policy exists for former employees.167
Thus, while the CIA, NSA, DIA, and ODNI require lifetime
prepublication review as a condition of employment, other agencies
appear to rely on various nondisclosure agreements, the conditions of
which may or may not require prepublication review after separation
from the agency.
C.
Issues and Problematic Patterns with Prepublication Review
1. Broad Discretion May Permit Opportunistic Post Hoc Classiﬁcation Decisions. — The lack of executive-branch-wide guidance allows agencies to
develop not just their own official policies but also their own uncodiﬁed
practices. Such discretion may lead agencies to exercise prepublication
review in an inconsistent and even opportunistic manner, particularly
with respect to former employees.
The experiences and opinions of CIA PRB officials ﬁll in the interstices left by broadly worded (and heavily redacted) policy documents
and highlight the scope of discretion granted to agencies. In a piece published in Studies in Intelligence, an internal magazine for the IC,168 John
Hollister Hedley discussed the purpose of prepublication review and his
role as Chairman of PRB at the time.169 It is important to note that this
piece was published in 1998, well before the reforms that took place in
2007 and during a time when the PRB was reviewing only the works of
former employees.170 Hedley begins by asserting that the “sole purpose of
prepublication review is to assist authors in avoiding inadvertent disclosure of classiﬁed information which, if disclosed, would be damaging to
national security—just that and nothing more.”171 Hedley acknowledges
that, “[n]otwithstanding a ﬁrm commitment to fairness and evenhandedness and with every intention of applying standard uniformly,” the
166. See About: Treasury Orders and Directives, U.S. Dep’t of Treasury, http://www.
treasury.gov/about/role-of-treasury/orders-directives/Pages/default.aspx (on ﬁle with the
Columbia Law Review) (last updated Sept. 15, 2013, 6:41 PM) (providing catalogue of all
Treasury orders and directives).
167. See John A. Cassara, Hide & Seek: Intelligence, Law Enforcement, and the
Stalled War on Terrorist Finance, at xiv (2006) (“Portions of this book were subject to
prepublication review and edited by the CIA, Department of State, and the U.S. Secret
Service. Postemployment publication policies by the Department of Homeland Security
and the Department of Treasury were also followed.” (emphasis added)).
168. The publication is classiﬁed, but unclassiﬁed extracts are available. See Center for
the Study of Intelligence, CIA, https://www.cia.gov/library/center-for-the-study-of-intellig
ence (on ﬁle with the Columbia Law Review) (last updated July 17, 2014, 4:06 PM).
169. Hedley, supra note 113.
170. See supra Part II.B.1 (discussing past PRB practices).
171. Hedley, supra note 113, at 75.
2015]
PREPUBLICATION REVIEW
441
Board’s deﬁnitions of “damage” and of “national security” are “neither
absolute nor constant.”172 He continues:
The important thing is for us to be reasonable and professional about what we protect. It does not take a genius to know
what information requires a hard look: for example, in an age
of terrorism and for privacy act considerations, we have to protect identities not already in the public domain. Also taboo—
because they impact adversely on our ability to conduct our
business, most of it necessarily in secret—are cover arrangements, liaison relationships, covert facilities, and unique collection and analytic capabilities. These constitute the sources and
methods that truly need protection . . . . In prepublication
reviews, we have to show we know the difference between what
truly is sensitive and what is not.173
While this provides insight into some factors the PRB might consider
when reviewing a manuscript, it also seems problematic in light of assertions by Hedley—and official policy—that the work of former employees
should only be reviewed for already classiﬁed (as opposed to “classiﬁable”)
material. His insistence that the PRB demonstrates it knows “the
difference between what truly is sensitive and what is not” implies that
the Board is not looking at whether the information has been classiﬁed,
but whether it should now be.
A more recent piece in Studies in Intelligence, originally published in
September 2011 and declassiﬁed in April 2013, gives an even more candid account of the review process and hints at the updated PRB regulations of 2011.174 The author (whose name is redacted) is the Directorate
of Intelligence’s representative to the PRB and appears to have been
there since 2007.175 The author begins with an analogy used throughout
the piece: The Board is to PRB regulations as Supreme Court Justices are
to the Constitution.176 Just as the Supreme Court “has interpreted the
Constitution in a variety of different rulings, often modifying the rulings
of previous justices” based on the “prevailing philosophy of the justices,”
the Board members interpret the PRB regulations “with the Agency’s
directors acting as a chief justice in setting the tone of interpretation.”177
The author then explains, as an example, that under Director Porter
Goss, the Board “tended to interpret the instruction to err on the side of
allowing very little to be published by CIA authors,” while more recent
directors, including George Tenet, Michael Hayden, and Leon Panetta,
172. Id. at 79.
173. Id. at 82–83.
174. See supra note 116 and accompanying text (explaining unavailability of CIA’s
2011 policies).
175. See CIA Prepublication Review in the Information Age, supra note 116, at 9
(describing “[author’s] current stint as the ﬁrst senior representative of the Directorate of
Intelligence” to PRB as sufficient to make previous roles seem “quaintly bucolic”).
176. Id. at 11.
177. Id.
442
COLUMBIA LAW REVIEW
[Vol. 115:417
“clearly favored a far looser interpretation to facilitate the publication of
a signiﬁcantly larger number of manuscripts.”178
The PRB has taken this analogy a step further and established “case
law precedents” in at least two contexts: ﬁctional works and works written
solely to fulﬁll an academic requirement.179 For ﬁctional works, the PRB
has developed “the James Bond literary genre test,” which applies a more
lenient standard to manuscripts that fall under the “spy novel” genre
while reserving the right of the Board to categorize the manuscript as
nonﬁction (and review it accordingly) should it determine that “the
tradecraft, operational details, or technology presented is very close to
reality.” 180 For academic papers, the PRB distinguishes seminar and
classroom-related papers from publishable theses and dissertations; the
former is “likely to receive PRB approval with the proviso that attempts to
publish it in any other forum would require a separate PRB review.”181
Notwithstanding such “case law precedents,” the author indicates that
the updated 2011 PRB regulations leave the Board free to deviate from
any previous decisions and evaluate each submission on a case-by-case
basis, considering “such things as the currency of the subject matter and
its relationship to a topic of public concern.”182
This quasi–case law approach may appear benign, particularly if one
believes that IC agencies, in light of their expertise, should be given some
deference in their interpretation of what will constitute damage to
national security. Since IC agencies, like any government organization,
have an interest in exercising control over their official messages, it is
reasonable to defer to the agency itself in determining the potential
adverse effects of disclosure by current employees.183 In terms of publications by former employees, however, the review—in light of the First
Amendment interests at stake—is meant to be less demanding and
limited to scrutiny for classiﬁed information.184 But the statements of CIA
officials involved in prepublication review seem to indicate that much
more is involved.185 In their statements, it is often unclear whether they
178. Id. Porter Goss’s emphasis on tightening scrutiny has been reported widely. See,
e.g., Scott Shane & Mark Mazzetti, Moves Signal Tighter Secrecy Within C.I.A., N.Y. Times
(Apr. 24, 2006), http://www.nytimes.com/2006/04/24/washington/24leak.html?pagewan
ted=print (on ﬁle with the Columbia Law Review) (quoting various named and unnamed
employees agreeing prepublication-review enforcement was noticeably stricter under new
director).
179. CIA Prepublication Review in the Information Age, supra note 116, at 17.
180. Id.
181. Id.
182. Id.
183. See supra note 130 and accompanying text (discussing CIA policies for current
employees).
184. See supra notes 64–65 and accompanying text (analyzing relevant case law).
185. In addition to the statements above, see also Declaration of Ralph S. Dimaio,
Information Review Officer, National Clandestine Service, Central Intelligence Agency at
10, Boening v. CIA, 579 F. Supp. 2d 166 (D.D.C. 2008) (No. 1:07CV00430 (EGS)),
2015]
PREPUBLICATION REVIEW
443
are commenting on the review process for current or former employees,
and this blurring seems to creep into the actual implementation of the
review process.186
It is also clear that the preference of the director—a political
appointee—has a discernible impact on the nature of review. Such inﬂuence is reﬂected strongly in the case of T.J. Waters, who sought to publish
a memoir based on his experiences as part of the ﬁrst class of CIA
recruits after 9/11.187 After receiving approval with minimal changes,188
he sold the approved portions of the manuscript to a publisher and
worked with the PRB to address the remaining areas.189 After several
months, the PRB returned with deletions that included “substantial portions of previously approved text that had undergone PRB/CIA review
and had been determined to be unclassiﬁed.” 190 In his suit, Waters
argued that this sudden reversal was due to the recent appointment of a
new director, Porter Goss,191 whose desire to reinstate a stricter culture of
secrecy has been widely reported.192 There is a broader debate within
administrative law about the role of politics in agency rulemaking,193 and
some may argue that it is only natural that a new director means new policy. But such an argument overlooks the First Amendment interests
involved and ignores relevant court decisions, which say that review
should be limited to scrutiny for classiﬁed information.194
available at http://fas.org/sgp/jud/boening/cia-dimaio.pdf (on ﬁle with the Columbia
Law Review) (opining review for classiﬁed material is “more art than science”).
186. See infra Part II.C.2 (discussing authors’ experiences with actual review process).
187. Memorandum in Support of Plaintiff’s Motion for Permanent Injunction or,
Alternatively, for Preliminary Injunction at 1–2, Waters v. CIA, No. 06-383 (RBW) (D.D.C.
ﬁled Mar. 7, 2006), available at http://fas.org/sgp/jud/waters-pimemo-030706.pdf (on
ﬁle with the Columbia Law Review).
188. Id. at 5 (“Only four words in the entire manuscript were determined to be
‘inappropriate for disclosure in the public domain and must be revised or deleted prior to
publication.’”).
189. Id. at 5–6.
190. Id. at 10.
191. See id. at 11–12 (“This new policy, which emanates from the CIA’s Director
Porter Goss, is intended to dissuade individuals to publish information, even if
unclassiﬁed, about their activities with the CIA.”).
192. See supra note 178 and accompanying text (describing Goss’s emphasis on strict
review of publications); see also Shane & Mazzetti, supra note 178 (reporting Goss sought
to “re-emphasize a culture of secrecy that has included a marked tightening of the review
process for books and articles by former agency employees”).
193. See, e.g., Mathew D. McCubbins, Roger G. Noll & Barry R. Weingast, Administrative Procedures as Instruments of Political Control, 3 J.L. Econ. & Org. 243, 246 (1987)
(“[T]he hypothesis we put forth is that much of administrative law . . . is written for the
purpose of helping elected politicians retain control of policymaking.”); Peter L. Strauss,
Overseer or “The Decider”? The President in Administrative Law, 75 Geo. Wash. L. Rev.
696, 704–05 (2007) (arguing in administrative law, “President’s role . . . is that of overseer
and not decider”).
194. See supra notes 64–65 and accompanying text (discussing Fourth and D.C.
Circuits’ opinions on scope of review).
444
COLUMBIA LAW REVIEW
[Vol. 115:417
Of course, agencies may classify information at any time,195 but the
fact that the CIA is conducting such “second looks” belies its insistence
that works by former employees are reviewed solely for properly classiﬁed
information. It is particularly problematic because authors undergoing
prepublication review are explicitly excluded from the right to appeal a
classiﬁcation decision to the Interagency Security Classiﬁcation Appeals
Panel (ISCAP),196 an interagency body that, among other things, considers challenges to classiﬁcation decisions by authorized holders of the
information197 and has been praised as one of the few truly effective
reform efforts. 198 Whatever the rationale in excluding authors from
ISCAP review, the effect is to make publication more onerous by depriving authors of one potential avenue of appeal.
Judicial review, which the Marchetti court held must be made available to any writer dissatisﬁed with the results of prepublication review,199
becomes all the more important under such conditions. If agencies are
permitted to conduct a “second look” and classify information post hoc,
the risk that they may do so to stiﬂe dissent or prevent embarrassment is
considerable. In McGehee v. Casey, the D.C. Circuit, in recognition of the
“strong ﬁrst amendment interest” of authors, held that reviewing courts
should “conduct a de novo review of the classiﬁcation decision, while giving deference to reasoned and detailed CIA explanations of that
classiﬁcation decision.” 200 Importantly, the court declared that review
should “go beyond the FOIA standard of review for cases reviewing CIA
censorship pursuant to secrecy agreements.”201 Similar to FOIA litigation,
however, courts have largely been unwilling to second-guess agencies’
classiﬁcation decisions. 202 Furthermore, the determination of whether
information was properly classiﬁed is conducted in camera and ex parte,
without the participation of the author’s counsel,203 denying authors the
195. See Exec. Order No. 13,526, 3 C.F.R. 298, 303 (2010) (permitting retroactive
classiﬁcation of material responsive to FOIA request).
196. Id. at 303.
197. See id. at 319–21 (establishing ISCAP and outlining its functions).
198. For a discussion of such praise, see infra note 262 and accompanying text.
199. See supra note 67 and accompanying text (discussing Marchetti court’s requirement of judicial review). The Supreme Court has upheld this idea. See Snepp v. United
States, 444 U.S. 507, 513 n.8 (1980) (per curiam) (recognizing review procedure
mandated by Snepp’s contract is subject to judicial review).
200. McGehee v. Casey, 718 F.2d 1137, 1148 (D.C. Cir. 1983).
201. Id.
202. The Marchetti opinion itself, though it predates McGehee and comes from a
different circuit, reﬂects this unwillingness. See United States v. Marchetti, 466 F.2d 1309,
1318 (4th Cir.), cert. denied, 409 U.S. 1063 (1972) (“The courts, of course, are illequipped to become sufficiently steeped in foreign intelligence matters to serve effectively
in the review of secrecy classiﬁcations in that area.”).
203. See Stillman v. CIA, 319 F.3d 546, 548–49 (D.C. Cir. 2003) (holding courts
should, before deciding constitutional question of whether author has First Amendment
right for his attorney to have access, inspect manuscript and any government pleadings to
2015]
PREPUBLICATION REVIEW
445
adversarial process afforded to other plaintiffs. Therefore, the agencies’
ability to classify post hoc, combined with the courts’ unwillingness or
inability (in light of the ex parte nature of the proceeding) to adequately
scrutinize classiﬁcation decisions, seems to create a considerable risk that
agencies may deny publication of information for political or other reasons with only tenuous connections to national security. The experiences
of prospective authors, discussed next, appear to conﬁrm this danger.
2. Experiences of Authors Suggest Discriminatory Enforcement. — The
experiences of authors undergoing prepublication review suggest that
review decisions may depend on whether the author is critical or supportive of the agency. It is impossible to inquire into the minds of officials
conducting review, and authors who have their works heavily redacted
may naturally allege—whether in earnest belief or disingenuously—that
some measure of bias was involved. But there is evidence that suggests
reviewing officials are looking for more than just classiﬁed information.
Tellingly, disputes over redactions in a work favorable to an agency are
almost nonexistent.204 There have been many memoirs praising the work
of IC agencies—some by high-level officials, including former CIA director George Tenet and the former director of the CIA’s National
Clandestine Service (NCS), Jose A. Rodriguez.205 But what distinguishes
the treatment is not the seniority of the official writing the memoir, but
the tone of the work. Indeed, former CIA director Stansﬁeld Turner, who
testiﬁed that Snepp’s book had caused “irreparable harm” to national
security,206 later found himself subject to the very treatment Snepp had
complained about—agency feedback Turner characterized as “irresponsible” and disputes over redactions that he described as ranging from
“borderline issues to the ridiculous.”207 In an apparent reversal from his
position during the Snepp dispute, Turner declared, “Clearly the Reagan
administration does not understand that oversight of intelligence in our
determine whether it can “resolve the classiﬁcation issue without the assistance of
plaintiff’s counsel”).
204. The only case involving a dispute over a work favorable to an agency—the
memoir by T.J. Waters, see supra notes 187–192 and accompanying text (recounting
review of memoir by Waters)—occurred during the early stages of Porter Goss’s
directorship. See Shane & Mazzetti, supra note 178 (quoting Waters’s attorney describing
book as “very positive” and one Waters thought “would be a great recruiting tool”). Goss
was known for his attempts to tighten personnel security generally. Supra notes 178, 191–
192 and accompanying text.
205. See infra notes 216–218 and accompanying text (discussing inconsistent levels of
review applied to works by high-level officials such as Tenet and Rodriguez).
206. See supra note 78 and accompanying text (referencing Turner’s testimony and
reliance on it in court).
207. Charles R. Babcock, Spy Agency Inﬁghting Hurt U.S., Turner Says, Wash. Post,
May 13, 1985, at A3, available at http://www.maebrussell.com/Stansfield%20Turner/
Stansfield%20Turner,%20Agency%20Infighting%20WP%205-13-85.html (on ﬁle with the
Columbia Law Review).
446
COLUMBIA LAW REVIEW
[Vol. 115:417
society includes constructive criticisms from constructive outsiders like
me.”208
Even high-level officials who author memoirs that praise the agency
and the IC may clash with prepublication-review boards. Former CIA
director Leon Panetta recently published a memoir described as “almost
unfailingly complimentary” toward the CIA,209 but nonetheless “clashed
with the [PRB] over [its] contents . . . and allowed his publisher to begin
editing and making copies of the book before he had received ﬁnal
approval from the CIA.210 Panetta became so frustrated with the process
that he appealed directly to the current CIA director, John Brennan, and
“threatened to proceed with publication without clearance from the
agency.”211 Such informal appeals to higher-level officials outside of the
PRB, combined with the fact that the CIA’s ﬁnal approval came well after
review copies were distributed, reeks of precisely the sort of favorable
treatment that gives credence to accusations that the prepublicationreview system is biased and discriminatory in its enforcement.212
Another recent high-proﬁle dispute involved former FBI agent Ali
Soufan and the PRB over publication of Soufan’s book, The Black Banners.
In it, Soufan criticizes both the CIA’s approach to interrogation of highproﬁle detainees in the period after 9/11 and its ineptitude prior to
9/11, claiming it withheld information that may have prevented the
attacks.213 Soufan submitted the book for prepublication review to both
the FBI and the CIA; it was cleared by the FBI with minor changes, but
the CIA demanded extensive cuts.214 After discussions with CIA officials,
Soufan decided not to litigate and instead published the book with the
black redaction bars in place to give readers an idea of how much he was
forced to withhold.215 He claims that much of what was taken out was
available publicly, including in the memoirs of other officials. For example, both the 9/11 Report and a memoir by former CIA Director George
208. Id.
209. Greg Miller, Panetta Clashed with CIA over Memoir, Tested Agency Review
Process, Wash. Post (Oct. 21, 2014), http://www.washingtonpost.com/world/national-sec
urity/panetta-clashed-with-cia-over-memoir-tested-agency-review-process/2014/10/21/6e6
a733a-5926-11e4-b812-38518ae74c67_story.html (on ﬁle with the Columbia Law Review).
210. Id.
211. Id.
212. See id. (“The CIA’s dispute with its former director, and its apparent decision not
to pursue the potential violation, could complicate the agency’s ability to negotiate with
other would-be authors and avoid accusations of favoritism.”).
213. Ali H. Soufan, The Black Banners: The Inside Story of 9/11 and the War Against
al-Qaeda 295–96 (2011).
214. Id. at xi.
215. See Greg Miller & Julie Tate, CIA Probes Publication Review Board over
Allegations of Selective Censorship, Wash. Post (May 31, 2012), http://articles.washington
post.com/2012-05-31/world/35455152_1_publications-review-board-harsh-interrogation-ci
a-critics (on ﬁle with the Columbia Law Review) (“[The book] was so heavily redacted that
he published [it] with black marks across many of its pages to show readers how much he
was forced to withhold.”).
2015]
PREPUBLICATION REVIEW
447
Tenet stated that the CIA had possession of the passport photo of one of
the 9/11 hijackers as early as January 2000.216 Soufan also alleges that
information included in a memoir by Jose A. Rodriguez, Jr., the former
head of the NCS, was redacted from his own book.217 Soufan claims that
credible sources inside the CIA had informed him that the agency did
not want the book published because it would prove embarrassing to the
agency.218
Others have similarly alleged that agencies disapproved of their
manuscripts solely because of their critical tones. Glenn Carle, a twentythree-year veteran of the CIA who conducted interrogations after 9/11,
published a memoir similarly criticizing the CIA’s interrogation policies.219 He submitted his work for review and, after more than a dozen
rewrites, elected to publish his book with the redactions in place, albeit
with the occasional caustic footnote.220 Speaking about his experience
working with the PRB, Carle stated, “Their goal was to intimidate me.
That was quite clear.”221 Because Carle happened to know members of
the board, he claimed they were exceptionally candid with him; one
allegedly asked him, “Don’t you realize that people could go to jail for
this?” in reference to certain passages in his book depicting interrogation
methods that Carle regarded as illegal.222
Another former CIA agent, writing under the pseudonym Ishmael
Jones, wrote a book critical of the CIA’s intelligence culture and submitted it for prepublication review.223 After over a year of negotiations, during the last six months of which he heard nothing from PRB, Jones
216. See Scott Shane, CIA Demands Cuts in Book About 9/11 and Terror Fight, N.Y.
Times (Aug. 25, 2011), http://www.nytimes.com/2011/08/26/us/26agent.html (on ﬁle
with the Columbia Law Review) (detailing redactions in Soufan’s book).
217. See Miller & Tate, supra note 215 (“‘Absolutely there are things that he was able
to talk about that were redacted from my book,’ Soufan said. ‘I think it has more to do
with trying to protect a narrative rather than protecting classiﬁed information.’”).
218. Id.; see also Benjamin Wittes, Is CIA Pre-Publication Review Biased?, Lawfare
(June 1, 2012, 7:09 AM), http://www.lawfareblog.com/2012/06/is-cia-pre-publicationreview-biased/ (on ﬁle with the Columbia Law Review) (expressing “sympathy for those who
suspect that one is allowed relative freedom to discuss the CIA [interrogation] program
only if one is defending it”).
219. See Glenn L. Carle, The Interrogator: An Education 291 (2011) (noting CIA
attempts to hide “rounded edges of wrongdoing, and obscure the corruption of our
institutions and of our systems of government caused by . . . coercive interrogation of
terrorists or terrorist suspects”).
220. See, e.g., id. at 60 n.1 (“Apparently the CIA fears that the redacted passage would
either humiliate the organization for incompetence or expose its officers to ridicule;
unless the Agency considers obtuse incompetence a secret intelligence method.”).
221. Miller, supra note 38.
222. Id.
223. Complaint at 5, United States v. Jones, No. 1:10-cv-00765-GBL-TRJ (E.D. Va. Apr.
18, 2012), available at http://fas.org/sgp/jud/jones/complaint.pdf (on ﬁle with the
Columbia Law Review).
448
COLUMBIA LAW REVIEW
[Vol. 115:417
elected to publish the book without approval.224 He claimed, “The Book
is highly critical of CIA management and outlines numerous instances of
waste, fraud, and abuse by the CIA, but contains no classiﬁed information.”225 Jones’s allegations appear credible enough that the CIA has
initiated an internal investigation into its review practices,226 the results of
which have not yet been reported.
Jones’s decision to publish without approval after months of delay
may also hint at the possibility of another troublesome development:
Would-be publishers, frustrated with their own experiences or having lost
faith in the process based on the complaints of others, may elect to forgo
the screening process altogether and publish unilaterally, effectively
becoming leakers. Such uncontrolled release deprives the executive of
any opportunity to review the materials and may lead to disclosures that
cause even more serious damage. One might expect such risk to encourage reform, but a recent memo by Director John Brennan indicates that
CIA review policies will only become more stringent.227
The use of such potentially abusive prepublication-review tactics is
not a recent phenomenon. In his book, Snepp describes the court battles
that led to the Supreme Court decision that bears his name,228 which
upheld the constitutionality of prepublication review.229 One allegation
among many stands out as particularly alarming: A named CIA lawyer
admitted to Snepp in private that they would have “ripped [Snepp’s]
manuscript to shreds on any pretext.”230
In a case predating Snepp’s experiences and the standardization of
nondisclosure agreements, the CIA sought to place a constructive trust
on the proceeds from a book Philip Agee, a former employee, published
without prepublication review and in violation of a secrecy agreement he
had signed.231 Agee freely admitted he violated the secrecy agreement
and that he intended to continue doing so as an author and journalist.232
As an affirmative defense, he argued that the CIA was practicing
224. Answer, Affirmative Defenses & Counterclaims of Defendant Ishmael Jones at 10–
13, Jones, No. 1:10-cv-00765-GBL-TRJ, available at http://fas.org/sgp/jud/jones/021111an
swer.pdf (on ﬁle with the Columbia Law Review).
225. Id. at 13.
226. Miller & Tate, supra note 215 (reporting initiation of internal investigation by
CIA).
227. See Kimberly Dozier, CIA Cracks Down on Its Own to Stop Leaks, AP: The Big
Story (June 26, 2013, 6:44 PM), http://bigstory.ap.org/article/cia-cracks-down-its-ownstop-leaks (on ﬁle with the Columbia Law Review) (reporting review of security policies
concluded “CIA also needs to be tougher with pre-publication review of articles or books
by former employees”).
228. Snepp v. United States, 444 U.S. 507 (1980) (per curiam).
229. See Snepp, Irreparable Harm, supra note 70, at 338–54 (describing events
leading up to and surrounding Snepp v. United States).
230. Id. at 346.
231. Agee v. CIA, 500 F. Supp. 506, 507 (D.D.C. 1980).
232. Id. at 508.
2015]
PREPUBLICATION REVIEW
449
discriminatory enforcement, choosing to bring suits only against authors
of works that were critical of the CIA.233 The court described this claim as
“substantial” and denied the government’s motion for summary judgment in light of evidence presented by Agee that “the CIA’s past enforcement record bears a considerable correlation with the agency’s
perception of the extent to which the material is favorable to the
agency.”234 To support claims of bias, Agee presented a list of ﬁve works
that were “critical of the Agency”; of those ﬁve, “four . . . spawned suits by
the Government to enforce the [prepublication-review] agreement,”
while no suits were ﬁled against authors whose works the agency did not
deem critical, despite those authors’ admitted failure to submit their
material for prepublication review.235
Former employees of other agencies have made similar allegations
that prepublication decisions are discriminatory.236 Such examples show
how, in the absence of rigid executive-branch-wide standards, agencies’
decisions can become biased.
3. Risk of Intimidation and Delay Tactics. — An analysis of authors’
experiences with prepublication review reveals a disturbing pattern that
borders on intimidation, but is at the very least an attempt to discourage
publication. Although the Marchetti court held that, in light of the First
Amendment rights at stake, agencies should respond to a review request
within a reasonable period of time (suggesting thirty days),237 agencies
have not adhered to this guideline.238 The opinion of Mark Zaid, an
attorney with extensive experience in prepublication review, is worth
quoting in its entirety:
233. Id.
234. Id. at 508–09.
235. See id. (pointing out authors whose works were not critical did not submit
manuscripts for review).
236. See, e.g., William J. Broad, Book Due Soon by Wen Ho Lee Is Causing Stir, N.Y.
Times (Aug. 5, 2001), http://www.nytimes.com/2001/08/05/us/book-due-soon-by-wenho-lee-is-causing-stir.html (on ﬁle with the Columbia Law Review) (describing experiences
of Wen Ho Lee and Danny Stillman, both former employees of Los Alamos who sought
publication); Jesselyn Radack, The Man the State Dept. Wants Silenced, Salon (Apr. 12,
2012, 9:26 AM), http://www.salon.com/2012/04/12/the_man_the_state_dept_wants_silen
ced/ (on ﬁle with the Columbia Law Review) (alleging former State Department employee
was required to preclear all social-media postings after publishing book critical of U.S.
reconstruction efforts in Iraq, while policy was not enforced against employees authoring
blogs “favorable” to Department); Leah Williams, FBI Attempts to Hold Sibel Edmond’s
Book Hostage, Whistle Blowers Protection Blog (Apr. 10, 2012), http://www.whistle
blowersblog.org/2012/04/articles/government-whistleblowers/fbi-whistleblowers/fbi-atte
mpts-to-hold-sibel-edmonds-book-hostage/ (on ﬁle with the Columbia Law Review) (describing excessive delay experienced by FBI whistleblower Sibel Edmonds in seeking publication of memoir).
237. See supra note 66 and accompanying text (discussing Marchetti opinion’s thirtyday requirement).
238. See supra note 157 (listing qualifying language used by agencies to avoid imposition of hard deadline).
450
COLUMBIA LAW REVIEW
[Vol. 115:417
Personally, I will openly concede that—notwithstanding the fact
that the PRB sends conﬂicting messages to its current and former employees regarding whether a speciﬁc deadline exists for
a response—a 30 day requirement is often unrealistic given the
manner in which the current process has been structured for
reviews. However, my experiences have revealed that delays that
extend one to two years before a ﬁnal response occurs have
become a common routine pattern and practice with the CIA.
The excessive delays have a signiﬁcant impact on the submitter,
especially since there is often a publication deadline involved or
an important public interest underlying the contents.239
Thus it appears that, even if a thirty-day response time is not feasible,
agencies are ignoring the precept of Marchetti and using undue delay as a
means to frustrate authors.
This practice of delay appears particularly nefarious in light of
another practice: As soon as a frustrated author brings suit under the
Administrative Procedures Act,240 the agency quickly issues a decision,
rendering the claim moot.241 Such cases appear to fall under a class of
agency action described as “capable of repetition yet evad[ing] review,”
entitling authors to an exception to the mootness doctrine, but courts
seem unwilling to ﬁnd such an exception.242 To discourage publication,
an agency could thus withhold a ﬁnal decision until an author makes the
decision to invest time and money to litigate the issue. Once litigation
commences, the APA claim is mooted by the agency’s rendering a decision. Those authors who instead choose to publish without approval are
sued to place a Snepp-style constructive trust on any proﬁts derived from
the book.243
239. Rule 56(f) Declaration of Mark S. Zaid, Esq. at 2, Boening v. CIA, 579 F. Supp. 2d
166 (D.D.C. 2008) (No. 07-430), available at http://fas.org/sgp/jud/boening/zaid111207
.pdf (on ﬁle with the Columbia Law Review).
240. 5 U.S.C. § 706(1) (2012) (“The reviewing court shall compel agency action
unlawfully withheld or unreasonably delayed . . . .”).
241. See, e.g., Berntsen v. CIA, 618 F. Supp. 2d 27, 29 n.2 (D.D.C. 2009) (“The
Amended Complaint also alleges that the CIA failed to timely complete its review of the
manuscript, and therefore, seeks an order to require the CIA to complete its review. This
allegation was subsequently mooted when the CIA completed its review of the
manuscript.”); Boening, 579 F. Supp. 2d at 172 (ﬁnding plaintiff’s claim moot, as PRB
issued decision after suit was ﬁled); Stillman v. CIA, 517 F. Supp. 2d 32, 36 (D.D.C. 2007)
(“Stillman’s APA claim is moot because there is no further relief that this Court can
provide as to that claim. Stillman has already received the ﬁnal classiﬁcation decision that
he sought from the defendant agencies.”).
242. See Boening, 579 F. Supp. 2d at 172 (“The capable of repetition doctrine applies
only in exceptional situations, and generally only where the named plaintiffs can make a
reasonable showing that he or she will again be subjected to the alleged illegality. Plaintiff
has made no such showing here.” (citation omitted)).
243. See, e.g., United States v. Jones, No. 1:10-cv-00765-GBL-TRJ, slip op. at 1 (E.D. Va.
Apr. 18, 2012), available at http://www.fas.org/sgp/jud/jones/041812-order.pdf (on ﬁle
with the Columbia Law Review) (granting government motion for constructive trust as
remedy for breach of nondisclosure agreement).
2015]
PREPUBLICATION REVIEW
451
Another disturbing pattern is more suggestive of intimidation. When
an agency ﬁnally renders an initial decision about required redactions,
those redactions are often so numerous that the work cannot realistically
be published.244 After the author ﬁles suit, however, the number of redactions falls dramatically.245 If the review board were truly scanning the
publication solely for classiﬁed information, it is unclear why the information should suddenly become unclassiﬁed once a suit has been ﬁled.
One possibility is that the initial redactions were not made in good faith
and were withdrawn only because the agency did not think it would withstand even a highly deferential level of judicial scrutiny. The redactions
could thus be interpreted as an attempt to intimidate the author and
discourage publication—an outsized number of redactions would render
the work unpublishable and require the author to appeal the agency
decision, causing even more delay.246
In the absence of transparent, executive-branch-wide guidance over
prepublication-review policies, the risk of inconsistency, bias, and undue
delay in agency review decisions is considerable. These risks boil down to
one overarching concern: IC agencies are abusing the prepublicationreview process to stiﬂe dissent while facilitating the promulgation of
works consistent with their own narrative. In light of the First
Amendment rights at stake,247 improvement is needed. It is important to
recognize that the Constitution grants the executive branch primary
authority to protect national security and conduct foreign affairs, but
improvements can be made without violating this basic tenet.
244. See, e.g., Reporter’s Transcript, Motions Hearing at 10, United States v. Jones,
No. 10-765 (E.D. Va. Apr. 18, 2012) [hereinafter Jones, Reporter’s Transcript], available at
http://www.fas.org/sgp/jud/jones/061511-hearing.pdf (on ﬁle with the Columbia Law
Review) (alleging PRB denied plaintiff right to publish anything but footnotes).
245. See, e.g., Berntsen, 618 F. Supp. 2d at 29 (“During the course of the litigation . . .
Berntsen provided the PRB with a classiﬁed submission identifying 97 items that he
wanted to publish in his manuscript. . . . [T]he PRB completed its review of the 97 items
and agreed to withdraw its objections as to all but 18 of the items . . . .”); Stillman, 517 F.
Supp. 2d at 35 (“In October 2000, Stillman was informed that the DOE, DoD, and CIA did
not want any part of his manuscript published. In June 2001, Stillman ﬁled a lawsuit . . .
challenging their classiﬁcation decision. Soon after Stillman ﬁled the lawsuit, the
government released the majority of the manuscript for publication.”); Plaintiff’s
Opposition to Defendants’ Second Motion for Summary Judgment at 34, Shaffer v. Def.
Intelligence Agency, 901 F. Supp. 2d 113 (D.D.C. 2013) (No. 10-2119(RMC)) [hereinafter
Shaffer, Plaintiff’s Opposition], available at http://fas.org/sgp/jud/shaffer/081213-opp.
pdf (on ﬁle with the Columbia Law Review) (“The defendants acknowledge that in 2010
they identiﬁed 433 particular passages for redaction based on alleged classiﬁcation. Not
even three years later, nearly half of the redacted passages no longer merited classiﬁcation.” (citation omitted)).
246. Indeed, one author’s attorney alleges, “In 2010, the DIA claimed that there
were . . . covert names of four operatives in Operation Dark Heart; an assertion [the author]
knew to be false . . . . ‘[W]hen [the author] confronted [the DIA reviewer] . . . on this
issue . . . he admitted “yeah—we just made that up as an excuse to stop publication.”’”
Shaffer, Plaintiff’s Opposition, supra note 245, at 35.
247. See supra Part I.B.1 (explaining First Amendment rights at stake).
452
COLUMBIA LAW REVIEW
[Vol. 115:417
III. SOLUTIONS
Part III discusses potential solutions to the issues identiﬁed in Part
II. Part III.A proposes a congressional solution—legislation that would
provide more speciﬁc, mandatory guidance for agencies conducting
review. Part III.B urges the executive to initiate its own reforms, most
importantly by establishing an independent, interagency review panel.
Finally, Part III.C discusses the need for robust judicial review of agency
decisions.
A.
Congressional Solutions
In recognition of authors’ First Amendment rights and the public
interest in access to information about the IC’s activities, Congress could
overhaul the prepublication-review process while still respecting the
executive’s Article II powers. As it stands, the only statutory authority for
prepublication review is the broadly worded mandate of the National
Security Act of 1947 to protect intelligence sources and methods.248
Congress could pass a law with mandatory guidelines for prepublication
review, addressing some of the issues identiﬁed in Part II.
First, Congress could mandate robust judicial review of disputed
classiﬁcation decisions. Such review could be similar in scope to that in
the proposed State Secrets Protection Act.249 That bill would have made
in camera inspection of classiﬁed information mandatory250 and avoided
ex parte proceedings by allowing attorneys to obtain security clearances
and participate in the inspection of classiﬁed evidence.251 The idea of
having to justify—in the presence of opposing counsel—a classiﬁcation
decision before a less deferential judge may encourage agencies to be
more judicious in their decisions to classify information,252 particularly if
the material is already in the public domain. It may even dissuade agencies from litigating the issue at all, in order to avoid attracting even more
attention to any redactions.
248. See supra notes 42–43 and accompanying text (analyzing text of statute).
249. S. 2533, 110th Cong. (2008).
250. See id. § 4052(b)(1)(A) (“[A]ll hearings under this chapter shall be conducted in
camera.” (emphasis added)).
251. See id. § 4052(b)(2) (permitting ex parte hearings only if, after in camera review
of evidence, court determines “interests of justice and national security cannot adequately
be protected”).
252. Ex parte proceedings have been criticized in other contexts; for example, after
the unauthorized disclosures by Edward Snowden, several proposals to reform the Foreign
Intelligence Surveillance Court were circulated, all of which would make the process more
adversarial. See, e.g., FISA Improvements Act of 2013, S. 1631, 113th Cong. § 4 (2013)
(authorizing Foreign Intelligence Surveillance Court to “appoint amicus curiae to assist
the court” in Foreign Intelligence Surveillance Act proceedings); FISA Court Reform Act
of 2013, S. 1467, 113th Cong. (2013) (establishing “Office of the Special Advocate” to
review applications and participate in proceedings before Foreign Intelligence
Surveillance Court).
2015]
PREPUBLICATION REVIEW
453
One could argue, however, that stronger judicial review would not
really inﬂuence agency action because there is no real drawback—even if
an agency’s classiﬁcation is invalidated by a judge, the agency is essentially in the same position it would have been in had it simply not classiﬁed the information. Because there is no penalty for an erroneous
classiﬁcation (other than the release of the information), there is no
incentive to classify information more carefully. The natural solution,
then, is to penalize agencies for erroneous classiﬁcations. Congress could
adopt a fee-shifting provision that would allow prevailing authors to collect attorneys’ fees and other reasonable costs. For classiﬁcation decisions
that suggest arbitrary or capricious action on the part of the agency,
courts could appoint a special counsel to determine whether disciplinary
action is warranted against the individuals involved. Such a mechanism
may seem excessively intrusive on the executive, but FOIA permits just
such penalties for agencies that wrongfully withhold information under
the statute.253 If such penalties are acceptable in the context of FOIA
litigation, where no fundamental rights are involved, then their adoption
in the context of prepublication review, where an individual’s First
Amendment rights are at stake, should not be any more contentious.
Second, Congress could specify that the works of former employees
are to be reviewed only for already-classiﬁed information, as opposed to
“classiﬁable” information. This would address concerns with post hoc
classiﬁcation of information for questionable purposes, assuming the
information—were it truly damaging—would have been classiﬁed before
the author’s work came before the agency. Of course, agencies may legitimately need to classify information post hoc, whether because of a purely
negligent failure to do so previously or because of a change in the
national-security environment. Congress should recognize this need
while preventing agency abuse of it. It can do so by mandating a
presumption for lower courts: Where an agency decides to classify information that was unclassiﬁed at the time the author submitted a manuscript, courts must presume that the classiﬁcation is illegitimate and shift
the burden to the agency to show a legitimate national-security interest.
Congress could go even further and mandate a higher burden of proof
for post hoc classiﬁcation—clear and convincing evidence, for example.
The Federal Rules of Criminal Procedure similarly encourage attorneys
to raise objections in a timely manner by mandating a prohibitively high
standard of review for objections raised only on appeal.254 This should
incentivize agencies to classify the information in the ﬁrst instance,
253. See 5 U.S.C. § 552(a)(4)(E) (2012) (permitting prevailing plaintiff to recover
“reasonable attorney fees and other litigation costs reasonably incurred”); id.
§ 552(a)(4)(F) (requiring special counsel to determine whether disciplinary action is
warranted where court questions whether agency personnel acted arbitrarily or
capriciously in withholding information).
254. Fed. R. Crim. P. 52(b) (permitting consideration only of “plain error” if issue was
not raised at trial).
454
COLUMBIA LAW REVIEW
[Vol. 115:417
rather than only after it has come before them under prepublication
review. Any concerns about depriving agencies of their ability to classify
information at any time255 should be allayed by the fact that this framework only calls for heightened judicial review of such post hoc classiﬁcation decisions, not an outright ban of them.
Of course, if post hoc classiﬁcations come under higher scrutiny,
burden shifting may perversely incentivize agencies to simply classify
everything ex ante, thereby exacerbating overclassiﬁcation in an already
overclassiﬁed intelligence community. This risk reiterates the importance
of the aforementioned penalties for erroneous classiﬁcation decisions.256
In order to deter agencies from resorting to overly broad ex ante
classiﬁcation, the potential penalties—both individual and institutional—
must be meaningful. There will certainly be close calls, where reasonable
minds disagree over the need to classify certain information, and the special counsel should take that into account when determining the need
for and extent of disciplinary proceedings. But the penalties must be
substantial enough to deter abuse and go beyond mere fee shifting. From
a policy perspective, such penalties would also help reduce classiﬁcation—a reduction that the executive readily acknowledges is necessary.257
These two mechanisms combined—penalties for both erroneous or
capricious classiﬁcations and a higher standard of proof for post hoc
classiﬁcations—should encourage agency officials to object only to truly
damaging national-security information during prepublication review
while preventing them from overclassifying information ex ante. But
both of these potential congressional solutions would ultimately rely on
the courts to heed the call for more robust judicial review—naïvely
optimistic in light of the judiciary’s track record,258 perhaps, but for
which there is some precedent.259
B.
Executive-Branch Solutions
In light of the wide discretion granted to it by Congress in the
National Security Act of 1947, 260 the executive branch could, itself,
standardize prepublication-review practices across agencies by establishing an interagency review panel. One example of such executive-branch255. See supra note 195 and accompanying text (explaining agencies may classify
information at any time).
256. See supra note 253 and accompanying text (discussing penalties for wrongful
withholding of information in response to FOIA request).
257. See, e.g., Reducing Over-Classiﬁcation Act, Pub. L. No. 111-258, 124 Stat. 2648
(2010) (codiﬁed as amended in scattered sections of 6 and 50 U.S.C.) (requiring
Department of Homeland Security to develop strategy to reduce overclassiﬁcation).
258. See supra note 202 and accompanying text (describing courts’ unwillingness to
second-guess classiﬁcation determinations).
259. See infra notes 279–289 and accompanying text (discussing case involving
unusually robust scrutiny of government’s classiﬁcation claims).
260. See supra notes 42–43 and accompanying text (discussing text of statute).
2015]
PREPUBLICATION REVIEW
455
wide oversight is ISCAP, the interagency panel that reviews challenges to
classiﬁcation decisions. 261 ISCAP has been praised as one of the few
successful reform efforts and should serve as a model going forward.262
Its success lies in its composition—it is made up of representatives from
“the major national security agencies that are the most proliﬁc classiﬁers,” including the Departments of Defense, Justice, and State, as well as
the CIA, the National Security Council, and the National Archives and
Records Administration.263 By placing declassiﬁcation authority beyond
the agency that classiﬁed the information originally and in ISCAP,
bureaucratic and political self-interest is removed from the equation.264
As a successful example of interagency review of classiﬁcation decisions, ISCAP review would seem to be a model for prepublication review.
As mentioned above, however, ISCAP review is not available to those
challenging a classiﬁcation determination in the context of prepublication review.265 The ﬁrst and most modest step toward reform, then, seems
simple enough: Permit dissatisﬁed authors to appeal their case to ISCAP,
an interagency panel with less interest in preventing publication of works
critical of one agency.
The executive branch should not stop there, however. Seizing on
ISCAP as a model, it could create, by executive order, a similar interagency panel to conduct prepublication review in the ﬁrst instance.
Creating an ISCAP-like interagency prepublication-review panel would
remove agency bias and address the “second look” concern articulated
above,266 as an interagency panel with no authority to classify information
could not classify information post hoc. Providing review in the ﬁrst
instance, as opposed to providing appellate review, would address allegations that agencies are using delay tactics to discourage potential authors
from publishing.267 Even if review were to take longer than the thirty days
recommended by Marchetti,268 authors would have no reason to suspect
that the delay was an attempt to discourage publication.
One potential criticism of such a panel is that agencies are in the
best position to determine the nature of damage that would result from
disclosure of information that they classify. Such concerns get to the
261. See supra notes 196–197 and accompanying text (describing ISCAP’s role).
262. See Steven Aftergood, Reducing Government Secrecy: Finding What Works, 27
Yale L. & Pol’y Rev. 399, 407 (2009) (praising ISCAP as “unexpectedly effective”).
263. Id.
264. Id. at 409.
265. See supra note 196 and accompanying text (identifying explicit exclusion of
prepublication-review decision from ISCAP review).
266. See supra notes 195–196 and accompanying text (describing issues with agencies
conducting “second looks” to determine whether information should be classiﬁed, as
opposed to whether it already actually is).
267. See, e.g., Jones, Reporter’s Transcript, supra note 244, at 10 (claiming eighteen
months of review constituted effective breach of secrecy agreement by agency).
268. See supra note 66 and accompanying text (explaining Marchetti thirty-day deadline is not intended to be hard-and-fast requirement).
456
COLUMBIA LAW REVIEW
[Vol. 115:417
larger debate over centralization versus decentralization in the context of
national security.269 While the agency may be in the best position to assess
the extent of the damage that would result, it is equally true that the U.S.
classiﬁcation system is meant to have executive-branch-wide application.270 Furthermore, a primary goal in the creation of ODNI was to
encourage collaboration and efficiency, 271 an area in which the
prepublication-review process could certainly improve. Most importantly,
an executive-branch-wide panel could better protect national security by
ensuring consistency. For example, an interagency panel could have prevented the confusion surrounding publication of Shaffer’s book, where
DoD and DIA had differing opinions of what was sensitive information.272
If the information divulged in the ﬁrst edition of Shaffer’s book truly was
damaging to national security, the executive should be encouraging the
creation of an interagency panel that could prevent such lapses.
Another potential issue is that of agency capture: Since the interagency panel would be made up of IC agencies, the review process would
not be independent. Instead of scrutinizing the information to determine whether it is actually sensitive, officials may simply defer to the
judgment of the agency whose information is being published with the
understanding that other officials will return the favor when their
agency’s information is on the line. While such misplaced esprit de corps
may indeed be difficult to prevent, the potential for bias should not be
fatal to the idea of an interagency panel. There are plenty of executivebranch bodies that command respect for their independence. The Office
of Legal Counsel, for example, has long been known for its strong,
independent legal analysis.273 If the executive could portray a stint as an
agency representative on the interagency review panel as prestigious, the
panel could attract talented, independent-minded employees and create
an atmosphere conducive to impartial review. While such a reputation
cannot be established overnight, the emphasis and prestige placed on
“joint duty assignments” since 9/11 show it would not be impossible.274
269. See generally, e.g., Anne Joseph O’Connell, The Architecture of Smart Intelligence: Structuring and Overseeing Agencies in the Post-9/11 World, 94 Calif. L. Rev.
1655, 1657 (2006) (questioning “siren call” of uniﬁcation of intelligence).
270. See supra notes 104–105 and accompanying text (describing executive-branchwide role of ISOO in classiﬁcation decisions).
271. See supra notes 107–109 and accompanying text (describing creation of ODNI).
272. See supra notes 1–2 and accompanying text (describing Shaffer’s experience with
review boards).
273. See, e.g., Trevor W. Morrison, Constitutional Alarmism, 124 Harv. L. Rev. 1688,
1708 (2011) (book review) (“OLC’s advice can and does impose meaningful legal constraints on its clients, and both OLC and its clients have powerful incentives to maintain
OLC’s reputation for doing so.”).
274. See IC Joint Duty, Office of the Dir. of Nat’l Intelligence, http://www.dni.gov/
index.php/about/organization/ic-joint-duty (on ﬁle with the Columbia Law Review) (last
visited Oct. 18, 2014) (quoting description of Joint Duty Program for IC professionals as
“‘key to improved national security’ and ‘innovative solution for improving cross-agency
understanding’”).
2015]
PREPUBLICATION REVIEW
457
The success of ISCAP is also encouraging, insofar as it suggests that the
culture of secrecy shared by the agencies does not necessarily create a
bond that prohibits objective, collective assessment.275 Furthermore, a
truly independent interagency panel may exude more legitimacy than
the prepublication offices of individual agencies. If so, authors would
probably be more willing to submit themselves to the process, which
would prevent them from becoming leakers and avoid all the extra damage concomitant with uncontrolled leaks.276
In broader terms, the proposal for an interagency panel is not very
radical. Separation-of-powers concerns are particularly important in the
context of national security,277 but none would exist here. Even if the
panel’s existence were mandated by legislation, the panel would be made
up of executive-branch agencies. It would not take power away from the
executive, but would instead promote consistency and efficiency in
national security. The executive branch should thus consider the
establishment of a centralized, interagency panel responsible for
conducting all prepublication review.
C.
Judicial Solutions
The most important but perhaps least promising source of solutions
for the problems with prepublication review remains: the judiciary.
Congress has already granted the judiciary the power and the means—
through in camera inspection and other tools—to question the executive’s classiﬁcation decisions in the context of FOIA, but courts seem
simply unwilling to do so.278 But there are some hints of promise. For
example, the D.C. District Court recently ruled against the FBI and in
favor of former agents attempting to publish, among other works, a memoir.279 The court’s opening line summarizes the case best:
This is a sad and discouraging tale about the determined
efforts of the FBI to censor various portions of a 500-page
manuscript, written by a former long-time FBI agent, severely
criticizing the FBI’s conduct of the investigation of a money
laundering scheme in which United States-based members of
the Hamas terrorist organization were using non-proﬁt
275. See supra notes 197–198, 262 and accompanying text (explaining ISCAP and its
success).
276. See supra Part II.C.2 (arguing would-be authors, frustrated with their experiences, may elect to publish unilaterally, depriving government of any opportunity to
review works and potentially leading to more damaging disclosures).
277. See, e.g., Morrison, supra note 273, at 1742–43 (criticizing Ackerman’s proposal
to create “Supreme Executive Tribunal” to settle disputes between Congress and executive
branch on separation-of-powers grounds).
278. See supra note 202 and accompanying text (describing unwillingness of courts to
scrutinize classiﬁcation decisions).
279. Wright v. FBI, 613 F. Supp. 2d 13, 24, 31 (D.D.C. 2009) (holding government had
not satisﬁed its burden for all but one censorship request).
458
COLUMBIA LAW REVIEW
[Vol. 115:417
organizations in this country to recruit and train terrorists and
fund terrorist activities both here and abroad.280
This opening did not bode well for the government, and the court proceeded to strike down all but one of the redactions proposed by the
FBI.281 But more remarkable is the way in which the court came to its
decision. Instead of simply taking the government at its word, the court
applied the Pickering balancing test282 to determine whether the government’s interest in censoring the material—which must be articulated
with “reasonable speciﬁcity”—outweighed the author’s First Amendment
interest in publication. 283 It also adopted the approach of the D.C.
Circuit,284 stating that the court’s review must be more searching than in
the FOIA context.285 In doing so, it granted some deference to the FBI,
but largely adopted a de novo approach.
For example, it swept aside the FBI’s argument that the freezing of
bank accounts constitutes “sensitive law enforcement activities, methods,
and capabilities” and declared it “common knowledge” that such techniques were used as part of the government’s counterterrorism strategy.286 Because it was “common knowledge,” the government could not
demonstrate that its interest in censoring that information outweighed
the public’s interest in disclosure; the government’s objection therefore
failed the Pickering test.287 The court also rejected censorship of material
that would otherwise have been protected by the internal-deliberativeprocess exemption of FOIA:288 The government failed in its burden to
“go beyond the FOIA standard” and present with “reasonable speciﬁcity
‘reasonably convincing and detailed evidence of a serious risk that intelligence sources and methods would be compromised’ by disclosure of the
materials discussed” such that it would outweigh the plaintiff’s First
280. Id. at 15.
281. See id. at 24 (“The Government Has Satisﬁed Its Burden to Justify Censorship for
Only One of Its Fourteen Objections to the Fatal Betrayals Manuscript.”).
282. See supra note 51 and accompanying text (articulating Pickering test).
283. Wright, 613 F. Supp. 2d at 22–23 (quoting post-Pickering case McGehee v. Casey,
718 F.2d 1137, 1148 (D.C. Cir. 1983)) (internal quotation marks omitted).
284. See supra note 201 and accompanying text (discussing D.C. Circuit’s understanding of appropriate scope of judicial review).
285. Wright, 613 F. Supp. 2d at 24 (“Consequently, censorship is prohibited under the
First Amendment where it fails the Pickering/NTEU balancing test, even if the material falls
within a FOIA Exemption.”).
286. Id. at 25–26. The court rejected the government’s contention that such techniques were covered by Exemption 7 of FOIA. Id.; see 5 U.S.C. § 552(b)(7) (2012)
(exempting from disclosure “records or information compiled for law enforcement purposes” that would “disclose techniques and procedures for law enforcement investigations
or prosecutions”).
287. Wright, 613 F. Supp. 2d at 25–26.
288. See 5 U.S.C. § 552(b)(5) (exempting from disclosure “inter-agency or intraagency memorandums or letters which would not be available by law to a party other than
an agency in litigation with the agency”).
2015]
PREPUBLICATION REVIEW
459
Amendment interests.289 Such unyielding judicial review provides a glimmer of hope, but it is still all too rare, despite recent decisions in the
FOIA context that may justify some enthusiasm.290
Some might argue that such robust review is rare because it should be
rare—both the courts and Congress have historically deferred to the
executive in the area of national security,291 and prepublication review is
no different. While courts should be mindful not to overstep their
authority—by reversing a classiﬁcation decision on a purely policy-based
rationale, for example—that argument ignores the First Amendment
rights involved. Where fundamental rights are involved, courts have
applied strict scrutiny even in the area of national security. 292
Congressional imprimatur—implied in FOIA and in this Note’s proposals,293 should they be adopted—further militates in favor of robust
review in the context of prepublication review.
The judiciary must protect the First Amendment rights of former
employees by acting as a meaningful check on the executive’s assertions
of secrecy. The Wright court showed that the judiciary has ample tools to
do just that, and other courts should follow its lead. If the solutions outlined above—including executive-branch reform—are to work, the judiciary must heed the call of the other branches and perform its duties. Only
then can the public be assured of a fair system of prepublication review
that protects the First Amendment rights of authors and guarantees
public access to their insights while simultaneously ensuring the public is
kept safe by preventing the disclosure of truly sensitive national security
information.
CONCLUSION
The highly secretive nature of the intelligence community makes it
difficult for the public to assess its performance. In this context, memoirs
and other publications by current and former employees provide crucial
insight. More importantly (and unlike in the FOIA context), a
fundamental constitutional right is at stake—the First Amendment right
289. Wright, 613 F. Supp. 2d at 29–30 (quoting McGehee v. Casey, 718 F.2d 1137, 1149
(D.C. Cir. 1983)).
290. See N.Y. Times Co. v. U.S. Dep’t of Justice, 756 F.3d 100, 124 (2d Cir. 2014)
(requiring OLC to release internal memorandum explaining legal reasoning as to
lawfulness of targeted killings of U.S. citizens by drone aircraft); ACLU v. CIA, 710 F.3d
422, 427–30 (D.C. Cir. 2013) (holding CIA could not refuse to conﬁrm or deny existence
of drone program in response to FOIA request in light of its public statements to
contrary).
291. See supra notes 19–24 and accompanying text (discussing deference granted to
executive in national-security matters).
292. See, e.g., Korematsu v. United States, 323 U.S. 214, 216 (1944) (applying strict
scrutiny to purported exercise of war powers even in time of national emergency).
293. See supra Part III.A (proposing Congressional solutions to prepublication
review).
460
COLUMBIA LAW REVIEW
[Vol. 115:417
of current and former employees to share their experiences working in
this highly secretive ﬁeld. Considering the important interests at issue for
both authors and the public, the current decentralized system of
prepublication review has proven inadequate in protecting such publications. The lack of executive-branch-wide guidance permits too much
discretion for individual agencies, creating the potential for abuse. In
light of the IC’s ever-expanding reach and inﬂuence and the disturbing
trend of overclassiﬁcation of national-security information, reasonable
access to all feasible means of evaluating the executive’s work in this area
is essential to maintaining an informed electorate. All three branches of
government must therefore take action to reform the prepublicationreview system and ensure it is not abused in a way that could stiﬂe dissent
or otherwise deprive the public of information necessary for evaluating
the IC’s performance and—ultimately—for participating effectively in
the democratic process.